WO2005116805A1

WO2005116805A1 - An interactive system and method

Info

Publication number: WO2005116805A1
Application number: PCT/SG2005/000144
Authority: WO
Inventors: Adrian David Cheok; Zhi Ying Zhou; Jiun Horng Pan
Original assignee: National University Of Singapore
Priority date: 2004-05-28
Filing date: 2005-05-09
Publication date: 2005-12-08
Also published as: US20050276444A1; US7474318B2

Abstract

An interactive system for interacting with a device in a mixed reality environment, the system comprising: an object having at least two surfaces, each surface having a marker; an image capturing device to capture images of the object in a first scene; and computer software to track the position and orientation of the object in the first scene by identifying a marker; wherein the computer software in response to manipulation of the object causes the device to perform an associated operation.

Description

Title An Interactive System and Method

Technical Field

The invention concerns an interactive system for interacting with a device in a mixed reality environment.

Background of the Invention

Relatively little change has occurred regarding user interfaces for computers. For decades, the standard input devices for a computer included a keyboard and mouse. Recent popular developments have included wireless keyboards and mice that communicate to a desktop terminal using Bluetooth or Radio Frequency. This eliminates the needs for cables, but requires the keyboard and mouse to use batteries. Another intuitive input method is voice recognition. This requires the computer to recognise and understand the voice of a user, and carry out a corresponding command. Voice recognition requires training the computer to recognise the speech patterns of a user. However, accuracy is still dependent on the processing power of the computer, the quality of the microphone and the clarity of the words spoken by the user.

These methods for interfacing with a computer cause user frustration as they are cumbersome and not immediately intuitive.

Summary of the Invention

In a first preferred aspect, there is provided an interactive system for interacting with a device in a mixed reality environment, the system comprising: an object having at least two surfaces, each surface having a marker; an image capturing device to capture images of the object in a first scene; and computer software to track the position and orientation of the object in the first scene by identifying a marker; wherein the computer software in response to manipulation of the object causes the device to perform an associated operation. To identify a marker for tracking the position and orientation of the object, at least one surface of the object may be tracked. The marker used for tracking the position and orientation of the object may be identified on a surface with the highest tracking confidence. The surface with the highest tracking confidence may be determined according to the extent of occlusion of its marker.

Advantageously, if the top surface of the object is occluded, the marker on the top surface is ascertainable and tracking of the object is possible by being able to identify a marker on another surface. This permits the user's intention as signified by the top surface of the object to be conveyed to the system, continuously and consistently.

In a second aspect, there is provided an interactive system for interacting with a device in a mixed reality environment, the system comprising: at least two objects, each object having at least two surfaces, each surface having a marker; an image capturing device to capture images of the objects in a first scene; and computer software to track the position and orientation of the objects in the first scene by identifying a marker on each object; wherein the computer software in response to manipulation of the objects and their arrangement relative to each other causes the device to perform an associated operation. In a third aspect, there is provided a method for interacting with a device in a mixed reality environment, the method comprising: capturing images of an object having at least two surfaces, each surface having a marker; and tracking the position and orientation of the object by identifying a marker; wherein in response to manipulation of the object, the device is made to perform an associated operation.

In a fourth aspect, there is provided a method for interacting with a device in a mixed reality environment, the method comprising: capturing images of at least two objects, each object having at least two surfaces, each surface having a marker; and tracking the position and orientation of the objects by identifying a marker on each object; wherein in response to manipulation of the objects and their arrangement relative to each other, the device is made to perform an associated operation.

The computer software may retrieve multimedia content associated with an identified marker, and generate a second scene including the associated multimedia content superimposed over the first scene in a relative position to the identified marker, to provide a mixed reality experience to a user.

The device may include a television, DVD player, lighting, or an air conditioner. Associated operations include power on or off, volume control, dimming level control or temperature control.

The device may include a computer. If the device is a computer, the computer software may cause other software applications on the computer to perform associated actions or tasks. Other software applications may include an MP3 player, a Media Player for playing video clips or movies, a Photo Album to display digital photos or an Internet web browser. For example, associated actions for a Media Player application include playing, pausing, fast forwarding or rewinding a video clip. In this example, translational movement of the object left or right instructs the Media Player to rewind or fast forward. Alternatively, rotating the object clockwise or anti-clockwise instructs the Media Player to rewind or fast forward.

Where there is more than one object, a first object may be used as an anchor for relative positioning of the associated multimedia content, and a second object may be used to operate the device or software application. The associated multimedia content may include a virtual representation of the device, or the user interface of the software application.

The computer software may rely on a state transition model to determine a response to a manipulation of the object, and to determine an associated operation for the device to perform.

Each physical action and physical property of the object may be virtually coupled to a virtual method and virtual attribute. Preferably, the marker includes a discontinuous border that has a single gap. Advantageously, the gap breaks the symmetry of the border and therefore increases the dissimilarity of the markers. More preferably, the marker comprises an image within the border. The image may be a geometrical pattern to facilitate template matching to identify the marker. The pattern may be matched to an exemplar stored in a repository of exemplars. Even more preferably, the colour of the border produces a high contrast to the background colour of the marker, to enable the background to be separated by the computer software. Advantageously, this lessens the adverse effects of varying lighting conditions.

The marker may be unoccluded to identify the marker.

The marker may be a predetermined shape. To identify the marker, at least a portion of the shape is recognised by the computer software. The computer software may determine the complete predetermined shape of the marker using the detected portion of the shape. For example, if the predetermined shape is a square, the computer software is able to determine that the marker is a square if one corner of the square is occluded.

The computer software may identify a marker if the border is partially occluded and if the pattern within the border is not occluded.

The interactive system may further comprise a display device such as a monitor, television screen or LCD, to display the second scene at the same time the second scene is generated. The display device may be a view finder of the image capture device or a projector to project images or video. The video frame rate of the display device may be in the range of twelve to thirty frames per second.

The image capture device may be mounted above the display device, and both the image capture device and display device may face the user. The object may be manipulated between the user and the display device.

Multimedia content may include 2D or 3D images, video and audio information.

Preferably, the at least two surfaces of the object are substantially planar. Preferably, the at least two surfaces are joined together. The object may be a cube or polyhedron. The object may be foldable, for example, a foldable cube for storytelling.

The computer software may be installed on a desktop or mobile computing device such as a Personal Digital Assistant (PDA), mobile telephone or other mobile communications device with a built-in computer processor.

The image capturing device may be a camera. The camera may be CCD or CMOS video camera.

The camera, computer software and display device may be provided in a single integrated unit.

The camera, computer software and display device may be located in remote locations.

The associated multimedia content may be superimposed over the first scene by rendering the associated multimedia content into the first scene, for every video frame to be displayed.

The position of the object may be calculated in three dimensional space A positional relationship may be estimated between the camera and the object. The camera image may be thresholded. Contiguous dark areas may be identified using a connected components algorithm. A contour seeking technique may identify the outline of these dark areas. Contours that do not contain four corners may be discarded. Contours that contain an area of the wrong size may be discarded.

Straight lines may be fitted to each side of the square contour. The intersections of the straight lines may be used as estimates of the corner positions. A projective transformation may be used to warp the region described by these corners to a standard shape. The standard shape may be cross-correlated with stored exemplars of markers to find the marker's identity and orientation. The positions of the marker corners may be used to identify a unique Euclidean transformation matrix relating to the camera position to the marker position.

The system may further comprise at least two objects, wherein the spatial relationship between the at least two objects is determined to cause a predetermined response from the multimedia content associated with the identified markers.

The spatial relationship may be selected from the group consisting of: distance, stacking and occlusion between the objects. The predetermined response may be selected from the group consisting of: interaction between the associated multimedia content, animation of at least one associated multimedia content and playback of an audio recording for at least one associated multimedia content.

In a fifth aspect, there is provided a software application for interacting with a device in a mixed reality environment, the application comprising: an image processing module to receive captured images of an object in a first scene from an image capturing device; and a tracking module to track the position and orientation of the object in the first scene by tracking at least two surfaces of the object where each surface has a marker, and identifying at least one marker; wherein the software application in response to manipulation of the object causes the device to perform an associated operation.

In a sixth aspect, there is provided an image capturing device for interacting with a second device in a mixed reality environment, the device comprising: an image capture module to capture images of an object in a first scene; and a tracking module to track the position and orientation of the object in the first scene by tracking at least two surfaces of the object where each surface has a marker, and identifying at least one marker; wherein in response to manipulation of the object, the second device is made to perform an associated operation.

In an seventh aspect, there is provided a computer program product comprised of a computer-readable medium for carrying computer-executable instructions for: receiving captured images of an object in a first scene from an image capturing device; and tracking the position and orientation of the object in the first scene by tracking at least two surfaces of the object where each surface has a marker, and identifying at least one marker; wherein in response to manipulation of the object, a device is made to perform an associated operation.

Brief Description of the Drawings

An example of the invention will now be described with reference to the accompanying drawings, in which: Figure 1 is a class diagram showing the 'abstraction' of graphical media and cubes of the interactive system; Figure 2 is a table showing the mapping of states and couplings defined in a

"method cube" of the interactive system;

Figure 3 is a table showing 'inheritance' in the interactive system;

Figure 4 is a table showing the virtual coupling in a 3D Magic Story Cube application; Figure 5 is a process flow diagram of the 3D Magic Story Cube application;

Figure 6 is a table showing the virtual couplings to add furniture in an Interior

Design application;

Figure 7 is a series of screenshots to illustrate how the 'picking up' and 'dropping off of virtual objects adds furniture to the board; Figure 8 is a series of screenshots to illustrate the method for re-arranging furniture;

Figure 9 is a table showing the virtual couplings to re-arrange furniture;

Figure 10 is a series of screenshots to illustrate 'picking up' and 'dropping off of virtual objects stacking furniture on the board; Figure 11 is a series of screenshots to illustrate throwing out furniture from the board;

Figure 12 is a series of screenshots to illustrate rearranging furniture collectively;

Figure 13 is a pictorial representation of the six markers used in the Interior Design application; Figure 14 is a class diagram illustrating abstraction and encapsulation of virtual and physical objects;

Figure 15 is a schematic diagram illustrating the coordinate system of tracking cubes;

Figure 16 is a process flow diagram of program flow of the Interior Design application;

Figure 17 is a process flow diagram for adding furniture;

Figure 18 is a process flow diagram for rearranging furniture; Figure 19 is a process flow diagram for deleting furniture; Figure 20 depicts a collision of furniture items in the Interior Design application; Figure 21 is a series of screenshots to illustrate interaction between virtual objects in response to the spatial relationship of the cubes; and Figure 22 is a screenshot from a 3D Vocabulary Book application.

Detailed Description of the Drawings

The drawings and the following discussion are intended to provide a brief, general description of a suitable computing environment in which the present invention may be implemented. Although not required, the invention will be described in the general context of computer-executable instructions, such as program modules, being executed by a personal computer. Generally, program modules include routines, programs, characters, components, data structures, that perform particular tasks or implement particular abstract data types. As those skilled in the art will appreciate, the invention may be practiced with other computer system configurations, including hand-held devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, console boxes, and the like. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Referring to Figure 1 , an interactive system is provided to allow interaction with a software application on a computer. In this example, the software application is a media player application for playing media files. Media files include AVI movie files or WAV audio files. The interactive system comprises software programmed using Visual C++ 6.0 on the Microsoft Windows XP platform, a computer monitor, and a Dragonfly Camera mounted above the monitor to track the desktop area.

Complex interactions using a simple Tangible User Interface (TUI) are enabled by applying Object Oriented Tangible User Interface (OOTUI) concepts to software development for the interactive system. The attributes and methods from objects of different classes are abstracted using Object Oriented Programming (OOP) techniques. Figure 1 at (a), shows the virtual objects (Image 10, Movie 11 , 3D Animated Object 12) structured in a hierarchical manner with their commonalities classified under the super class, Graphical Media 13. The three subclasses that correspond to the virtual objects are Image 10, Movie 11 and 3D Animated Object 12. These subclasses inherit attributes and methods from the Graphical Media super class 13. The Movie 11 and 3D Animated Object 12 subclasses contain attributes and methods that are unique to their own class. These attributes and methods are coupled with physical properties and actions of the TUI decided by the state of the TUI. Related audio information can be associated with the graphical media 11, 12, 13, such as sound effects. In the system, the TUI allows control of activities including searching a database of files and sizing, scaling and moving of graphical media 11 , 12, 13. For movies and 3D objects 11 , 12, activities include playing/pausing, fast-forwarding and rewinding media files. Also, the sound volume is adjustable.

In this example, the TUI is a cube. A cube in contrast to a ball or complex shapes, has stable physical equilibriums on one of its surfaces making it relatively easier to track or sense. In this system, the states of the cube are defined by these physical equilibriums. Also, cubes can be piled on top of one another. When piled, the cubes form a compact and stable physical structure. This reduces scatter on the interactive workspace. Cubes are intuitive and simple objects familiar to most people since childhood. A cube can be grasped which allows people to take advantage of keen spatial reasoning and leverages off prehensile behaviours for physical object manipulations.

The position and movement of the cubes are detected using a vision-based tracking algorithm to manipulate graphical media via the media player application. Six different markers are present on the cube, one marker per surface. In other instances, more than one marker can be placed on a surface. The position of each marker relative to each another is known and fixed because the relationship of the surfaces of the cube is known. To identify the position of the cube, any one of the six markers is tracked. This ensures continuous tracking even when a hand or both hands occlude different parts of the cube during interaction. This means that the cubes can be intuitively and directly handled with minimal constraints on the ability to manipulate the cube.

The state of artefact is used to switch the coupling relationship with the classes. The states of each cube are defined from the six physical equilibriums of a cube, when the cube is resting on any one of its faces. For interacting with the media player application, only three classes need to be dealt with. A single cube provides adequate couplings with the three classes, as a cube has six states. This cube is referred to as an "Object Cube" 14. However, for handling the virtual attributes/methods 17 of a virtual object, a single cube is insufficient as the maximum number of couplings has already reached six, for the Movie 11 and 3D Animated object 12 classes. The total number of couplings is six states of a cube < 3 classes + 6 attributes/methods 17. This exceeds the limit for a single cube. Therefore, a second cube is provided for coupling the virtual attribute/methods 17 of a virtual object. This cube is referred to as a "Method Cube" 15.

The state of the "Object Cube" 14 decides the class of object displayed and the class with which the "Method Cube" 15 is coupled. The state of the "Method Cube" 15 decides which virtual attribute/method 17 the physical property/action 18 is coupled with. Relevant information is structured and categorized for the virtual objects and also for the cubes. Figure 1, at (b) shows the structure of the cube 16 after abstraction.

The "Object Cube" 14 serves as a database housing graphical media. There are three valid states of the cube. When the top face of the cube is tracked and corresponds to one of the three pre-defined markers, it only allows displaying the instance of the class it has inherited from, that is the type of media file in this example. When the cube is rotated or translated, the graphical virtual object is displayed as though it was attached on the top face of the cube. It is also possible to introduce some elasticity for the attachment between the virtual object and physical cube. These states of the cube also decide the coupled class of "Method Cube" 15, activating or deactivating the couplings to the actions according to the inherited class.

For elasticity, after a marker is tracked last at position A, the system may lose tracking of the marker due to the occlusion caused by the user's hands. When the marker is re-tracked later at position B, the virtual object is displayed in the subsequent frames as it is bounced from position A to position B. This enables a smooth transition to avoid the flashing of the object display when the system loses tracking of the marker or object with markers. Referring to Figure 2, on the 'Method Cube' 15, the properties/actions 18 of the cube are respectively mapped to the attributes/methods 17 of the three classes of the virtual object. Although there are three different classes of virtual object which have different attributes and methods, new interfaces do not have to be designed for all of them. Instead, redundancy is reduced by grouping similar methods/properties and implementing the similar methods/properties using the same interface.

In Figure 2, methods 'Select' 19, 'Scale X-Y' 20 and 'Translate' 21 are inherited from the Graphical Media super-class 13. They can be grouped together for control by the same interface. Methods 'Set Play/Stop' 23, 'Set Animate/Stop', 'Adjust Volume' 24 and 'Set Frame Position' 22 are methods exclusive to the individual classes and differ in implementation. Although the methods 17 differ in implementation, methods 17 encompassing a similar idea or concept can still be grouped under one interface. As shown, only one set of physical property/action 18 is used to couple with the 'Scale' method 20 which all three classes have in common. This is an implementation of polymorphism in OOTUI. This is a compact and efficient way of creating TUIs by preventing duplication of interfaces or information across classifiable classes and the number of interfaces in the system is reduced. Using this methodology, the number of interfaces is reduced from fifteen (methods for image - three interfaces, movie - six interfaces, 3D object - six interfaces) to six interfaces. This allows the system to be handled by six states of a single cube.

Referring to Figure 3, the first row of pictures 30 shows that the cubes inherit properties for coupling with methods 31 from 'movie' class 11. The user is able to toggle through the scenes using the 'Set Frame Method' 32 which is in the inherited class. The second row 35 shows the user doing the same task for the '3D object' class 12. The first picture in the third row 36 shows that 'image' class 10 does not inherit the 'Set Frame Method' 32 hence a red cross appears on the surface. The second picture shows that the 'Object Cube' 14 is in an undefined state indicated by a red cross.

The rotating action of the 'Method Cube' 15 to the 'Set Frame' 32 method of the movie 11 and animated object 12 is an intuitive interface for watching movies. This method indirectly fulfils functions on a typical video-player such as 'fast-forward' and 'rewind'. Also, the 'Method Cube' 15 allows users to 'play/pause' the animation.

The user can size graphical media of all the three classes by the same action, that is, by rotating the 'Method Cube' 15 with "+" as the top face (state 2). This invokes the 'Size' method 20 which changes the size of the graphical media with reference to the angle of the cube to the normal of its top face. From the perspective of a designer of TUIs, the 'Size' method 20 is implemented differently for the three classes 10, 11 , 12. However, this difference in implementation is not perceived by the user and is transparent.

To enhance the audio and visual experience for the users, visual and audio effects are added to create an emotionally evocative experience. For example, an animated green circular arrow and a red cross are used to indicate available actions. Audio feedback includes a sound effect to indicate state changes for both the object and method cubes.

Example - 3D Vocabulary Book

Referring to Figure 22, a 3D interactive vocabulary book for children is an application of the interactive system. The 3D interactive vocabulary book requires interaction from two cubes. The "object cube" on the left of the screenshot has six surfaces. Each surface represents a category of 3D objects for children to learn. Figure 22 shows a "vehicle" category. The "method cube" on the right of the screenshot is used to navigate the "vehicle" database. When the user rotates the "method cube" according to the navigation pattern shown above the top face of the cube, the 3D model shown above the "object cube" is changed from a tank to a car. A pop-up 2D text displays the word "tank" in different languages including a brief description. The model may be animated. If the model is animated, an engine noise is played together with a human narration of the brief description. Different pronunciations of the word in different languages may also be played. Again, the user is provided with other interactions including resizing and moving objects.

Example - 3D Magic Story Cube application

Another application of the interactive system is the 3D Magic Story Cube application. In this application, the story cube tells a famous Bible story, "Noah's Ark". Hardware required by the application includes a computer, a camera and a foldable cube. Minimum requirements for the computer are at least of 512MB RAM and a 128MB graphics card. In one example, an IEEE 1394 camera is used. An IEEE 1394 card is installed in the computer to interface with the IEEE 1394 camera. Two suitable IEEE 1394 cameras for this application are the Dragonfly cameras or the Firefly cameras. Both of these cameras are able to grab color images at a resolution of 640x480 pixels, at a speed of 30Hz. This is able to view the 3D version of the story whilst exploring the folding tangible cube. The higher the capture speed of the camera is, the more realistic the mixed reality experience is to the user due to a reduction in latency. The higher the resolution of the camera, the greater the image detail, thus improving tracking accuracy. A foldable cube is used as the TUI for 3D storytelling. Users can unfold the cube in a unilateral manner. Foldable cubes have previously been used for 2D storytelling with the pictures printed out on the cube's surfaces.

The software and software libraries used in this application are Microsoft Visual C++ 6.0, DirectX, OpenGL, GLUT and MXR Development toolkit. Microsoft Visual C++ 6.0 is used as the development tool. It features a fully integrated editor, compiler, and debugger to make coding and software development easier. Libraries for other components are also integrated. In Virtual Reality (VR) mode, DirectX, OpenGL and GLUT play important roles for graphics display. OpenGL is the premier environment for developing portable, interactive 2D and 3D graphics applications. OpenGL is responsible for all the manipulation of the graphics in 2D and 3D in VR mode. GLUT is the OpenGL Utility Toolkit and is a window system independent toolkit for writing OpenGL programs. It is used to implement a windowing application programming interface (API) for OpenGL. The MXR Development Toolkit enables developers to create Augmented Reality (AR) software applications. It is used for programming the applications mainly in video capturing and marker recognition. The MXR Toolkit is a computer vision tool to track fiducials and to recognise patterns within the fiducials. The use of a cube with a unique marker on each face allows for the position of the cube to be tracked by the computer by the MXR Toolkit continuously.

Referring to Figure 4, the 3D Magic Story Cube application applies a simple state transition model 40 for interactive storytelling. Appropriate segments of audio and 3D animation are played in a pre-defined sequence when the user unfolds the cube into a specific physical state 41. The state transition is invoked only when the contents of the current state have been played. Applying OOTUI concepts, the virtual coupling of each state of the foldable cube can be mapped 42 to a page of digital animation.

Referring to Figure 5, an algorithm 50 is designed to track the foldable cube that has a different marker on each unfolded page. The relative position of the markers is tracked 51 and recorded 52. This algorithm ensures continuous tracking and determines when a page has been played once through. This allows the story to be explored in a unidirectional manner allowing the story to maintain a continuous narrative progression. When all the pages of the story have played through once, the user can return to any page of the story to watch the scene play again.

A few design considerations that are kept in mind when designing the system is the robustness of the system during bad lighting conditions and the image resolution.

The unfolding of the cube is unidirectional allowing a new page of the story to be revealed each time the cube is unfolded. Users can view both the story illustrated on the cube in its non-augmented view (2D view) and also in its augmented view (3D view). The scenarios of the story are 3D graphics augmented on the surfaces of the cube.

The AR narrative provides an attractive and understandable experience by introducing 3D graphics and sound in addition to 3D manipulation and 3D sense of touch. The user is able to enjoy a participative and exploratory role in experiencing the story. Physical cubes offer the sense of touch and physical interaction which allows natural and intuitive interaction. For example, the user can move a control cube close to the story cube to remove or add in a new story character or story object. Also, the physical cubes allow social storytelling between an audience as they naturally interact with each other.

To enhance user interaction and intuitiveness of unfolding the cube, animated arrows appear to indicate the direction of unfolding the cube after each page or segment of the story is played. Also, the 3D virtual models used have a slight transparency of 96% to ensure that the user's hands are still partially visible to allow for visual feedback on how to manipulate the cube.

The rendering of each page of the story cube is carried out when one particular marker is tracked. As the marker can be small, it is also possible to have multiple markers on one page. Since multiple markers are located on the same surface in a known layout, tracking one of the markers ensures that the positions of the other markers are known. This is a performance issue to facilitate more robust tracking. To assist with synchronisation, the computer system clock is used to increment the various counters used in the program. This causes the program to run at varying speeds for different computers. An alternative is to use a constant frame rates method in which a constant number of frames are rendered every second. To achieve constant frame rates, one second is divided in many equal sized time slices and the rendering of each frame starts at the beginning of each time slice. The application has to ensure that the rendering of each frame takes no longer than one time slice, otherwise the constant frequency of frames will be broken. To calculate the maximum possible frame rate for the rendering of the 3D Magic Story Cube application, the amount of time needed to render the most complex scene is measured. From this measurement, the number of frames per second is calculated.

Example - Interior Design Application

A further application developed for the interactive system is the Interior Design application. In this application, the MXR Toolkit is used in conjunction with a furniture board to display the position of the room by using a book as a furniture catalogue.

MXR Toolkit provides the positions of each marker but does not provide information on the commands for interacting with the virtual object. The cubes are graspable allowing the user to have a more representative feel of the virtual object. As the cube is graspable (in contrast to wielding a handle), the freedom of movement is less constrained. The cube is tracked as an object consisting of six joined markers with a known relationship. This ensures continual tracking of the cube even when one marker is occluded or covered.

In addition to cubes, the furniture board has six markers. It is possible to use only one marker on the furniture board to obtain a satisfactory level of tracking accuracy. However, using multiple fiducials enables robust tracking so long as one fiducial is not occluded. This is crucial for the continuous tracking of the cube and the board. To select a particular furniture item, the user uses a furniture catalogue or book with one marker on each page. This concept is similar to the 3D Magic Story Cube application described. The user places the cube in the loading area beside the marker which represents a category of furniture of selection to view the furniture in AR mode.

Referring to Figure 14, prior to determining the tasks to be carried out using cubes, applying OOTUI allows a software developer to deal with complex interfaces. First, the virtual objects of interest and their attributes and methods are determined. The virtual objects are categorized into two groups: stackable objects 140 and unstackable objects 141. Stackable objects 140 are objects that can be placed on top of other objects, such as plants, TVs and Hi-Fi units. They can also be placed on the ground. Both groups 140, 141 inherit attributes and methods from their parent class, 3D Furniture 142. Stackable objects 140 have an extra attribute 143 of its relational position with respect to the object it is placed on. The result of this abstraction is shown in Figure 14 at (a).

For virtual tool cubes 144, the six equilibriums of the cube are defined as one of the factors determining the states. There are a few additional attributes to this cube to be used in complement with a furniture catalogue and a board. Hence, we have a few additional attributes such as relational position of a cube with respect to the book 145 and board 146. These additional attributes coupled with the attributes inherited from the Cube parent class 144 determines the various states of the cube. This is shown in Figure 14 at (b).

To pick up an object intuitively, the following is required:

1) Move into close proximity to a desired object

2) Make a 'picking up' gesture using the cube

The object being picked up will follow that of the hand until it is dropped. When a real object is dropped, we expect the following:

1) Object starts dropping only when hand makes a dropping gesture

2) In accordance with the laws of gravity, the dropped object falls directly below that of the position of the object before it is dropped 3) If the object is dropped at an angle, it will appear to be at an angle after it is dropped. These are the underlying principles governing the adding of a virtual object in Augmented Reality.

Referring to Figure 6, applying OOTUI, the couplings 60 are formed between the physical world 61 and virtual world 62 for adding furniture. The concept of translating 63 the cube is used for other methods such as deleting and rearranging furniture. Similar mappings are made for the other faces of the cube.

To determine the relationship of the cube with respect to the book and the board, the position and proximity of the cubes with respect to the virtual object need to be found. Using the MXR Toolkit, co-ordinates of each marker with respect to the camera is known. Using this information, matrix calculations are performed to find the proximity and relative position of the cube with respect to other items including the book and board.

Figure 7 shows a detailed continuous strip of screenshots to illustrate how the 'picking up' 70 and 'dropping off 71 of virtual objects adds furniture 72 to the board.

Referring to Figure 8, similar to adding a furniture item, the idea of 'picking up' 80 and dropping off is also used for rearranging furniture. The "right turn arrow" marker 81 is used as the top face as it symbolises moving in all directions possible in contrast to the "+" marker which symbolises adding. Figure 9 shows the virtual couplings to re-arrange furniture.

When designing the AR system, the physical constraints of virtual objects are represented as objects in reality. When introducing furniture in a room, there is a physical constraint when moving the desired virtual furniture in the room. If there is a virtual furniture item already in that position, the user is not allowed to 'drop off another furniture item in that position. The nearest position the user can drop the furniture item is directly adjacent the existing furniture item on board.

Referring to Figure 10, a smaller virtual furniture item can be stacked on to larger items. For example, items such as plants and television sets can be placed on top of shelves and tables as well as on the ground. Likewise, items placed on the ground can be re-arranged to be stacked on top of another item. Figure 10 shows a plant picked up from the ground and placed on the top of a shelf. Referring to Figure 11, to delete or throw out an object intuitively, the following is required: 1) Go to close proximity to desired object 110; 2) Make a 'picking up' gesture using the cube 111 ; and 3) Make a flinging motion with the hand 112;

Referring to Figure 12, certain furniture items can be stacked on other furniture items. This establishes a grouped and collective relationship 120 with certain virtual objects. Figure 12 shows the use of the big cube (for grouped objects) in the task of rearranging furniture collectively.

Visual and audio feedback are added to increase intuitiveness for the user. This enhances the user experience and also effectively utilises the user's sense of touch, sound and sight. Various sounds are added when different events take place. These events include selecting a furniture object, picking up, adding, rearranging and deleting. Also, when a furniture item has collided with another object on the board, an incessant beep is continuously played until the user moves the furniture item to a new position. This makes the augmented tangible user interface more intuitive since providing both visual and audio feedback increases the interaction with the user.

The hardware used in the interior design application includes the furniture board and the cubes. The interior design application extends single marker tracking described earlier. The furniture board is two dimensional whereas the cube is three dimensional for tracking of multiple objects.

Referring to Figure 13, the method for tracking user ID cards is extended for tracking the shared whiteboard card 130. Six markers 131 are used to track the position of the board 130 so as to increase robustness of the system. The transformation matrix for multiple markers 131 is estimated from visible markers so errors are introduced when fewer markers are available. Each marker 131 has a unique pattern 132 in its interior that enables the system to identify markers 131 , which should be horizontally or vertically aligned and can estimate the board rotation. The showroom is rendered with respect to the calculated centre 133 of the board. When a specific marker above is being tracked, the centre 133 of the board is calculated using some simple translations using the preset X-displacement and Y- displacement. These calculated centres 133 are then averaged depending on the number of markers 131 tracked. This ensures continuous tracking and rendering of the furniture showroom, on the board 130 as long as one marker 131 is being tracked.

When the surface of the marker 131 is approaching parallel to the line of sight, the tracking becomes more difficult as less pixels are used for recognition. When the marker flips over, the tracking is lost. Since the whole area of the marker 131 must always visible to ensure a successful tracking, it does not allow any occlusions on the marker 131. This leads to the difficulties of manipulation and natural two- handed interaction.

Referring to Figure 15, one advantage of this algorithm is that it enables direct manipulation of cubes with both hands. When one hand is used to manipulate the cube, the cube is always tracked as long as at least one of the six faces of the cube is detected. The algorithm used to track the cube is as follows: 1. Detect all the surface markers 150 and calculate the corresponding transformation matrix (Tcm) for each detected surface.

2. Choose a surface with the highest tracking confidence and identify its surface ID, that is top, bottom, left, right, front, and back.

3. Calculate the transformation matrix from the marker co-ordinate system to the object co-ordinate system (Tmo) 151 based on the physical relationship of the chosen marker and the cube.

4. The transformation matrix from the object co-ordinate system 151 to the camera co-ordinate system (Tco) 152 is calculated by: Tco = Tcm^"1 X Tmo.

Figure 16 shows the execution of the AR Interior Design application in which the board 160, small cube 161 and big cube 162 are concurrently being searched for.

To enable the user to pick up a virtual object when the cube is near the marker 131 of the furniture catalogue requires the relative distance between the cube and the virtual object to be known. Since the MXR Toolkit returns the camera co-ordinates of each marker 131 , markers are used to calculate distance. Distance between the marker on the cube and the marker for a virtual object is used for finding the proximity of the cube with respect to the marker.

The camera co-ordinates of each marker can be found. This means that the camera co-ordinates of the marker on the cube and that of the marker of the virtual object is provided by the MXR Toolkit. In other words, the co-ordinates of the cube marker with respect to the camera and the co-ordinates of the virtual object marker is known. TA is the transformation matrix to get from the camera origin to the virtual object marker. TB is the transformation matrix to get from the camera origin to the cube marker. However this does not give the relationship between cube marker and virtual object marker. From the co-ordinates, the effective distance can be found.

By finding TA -1 , the transformation matrix to get from the virtual object to the camera origin is obtained. Using this information, the relative position of cube with respect to virtual object marker is obtained. The proximity of the cube and the virtual object is of interest only. Hence only the translation needed to get from the virtual object to the cube is required (i.e. Tx, Ty, Tz), and the rotation components can be ignored.

Tz is used to measure if the cube if it is placed on the book or board. This sets the stage for picking and dropping objects. This value corresponds to the height of the cube with reference to the marker on top of the cube. However, a certain range around the height of the cube is allowed to account for imprecision in tracking.

Tx, Ty is used to determine if the cube is within a certain range of the book or the board. This allows for the cube to be in an 'adding' mode if it is near the book and on the loading area. If it is within the perimeter of the board or within a certain radius from the centre of the board, this allows the cube to be re-arranged, deleted, added or stacked onto other objects. There are a few parameters to determine the state of the cube, which include: the top face of the cube, the height of the cube, and the position of the cube with respect to the board and book.

The system is calibrated by an initialisation step to enable the top face of the cube to be determined during interaction and manipulation of the cube. This step involves capturing the normal of the table before starting when the cube is placed on the table. Thus, the top face of the cube can be determined when it is being manipulated above the table by comparing the normal of the cube and the table top. The transformation matrix of the cube is captured into a matrix called tfmTable. The transformation matrix encompasses all the information about the position and orientation of the marker relative to the camera. In precise terms, it is the Euclidean transformation matrix which transforms points in the frame of reference of the tracking frame, to points in the frame of reference in the camera. The full structure in the program is defined as:

The last row in equation 1 is omitted as it does not affect the desired calculations. The first nine elements form a 3x3 rotation matrix and describe the orientation of the object. To determine the top face of the cube, the transformation matrix obtained from tracking each of the face is used and works out the following equation. The transformation matrix for each face of the cube is called tfmCube.

Dot_product = tfmCube.ri₃* tfmTable.r^ + tfmCube.r₂3* tfmTables₂3 + tfmCube.Xzz* tfmTable. r₃₃ (Equation 2)

The face of the cube which produces the largest Dot_product using the transformation matrix in equation 2 is determined as the top face of the cube. There are also considerations of where the cube is with respect to the book and board. Four positional states of the cube are defined as - Onboard, Offboard, Onbook and Offbook. The relationship of the states of cube with the position of it, is provided below:

Referring to Figure 17, adding the furniture is done by using "+" marker as the top face of the cube 70. This is brought near the furniture catalogue with the page of the desired furniture facing up. When the cube is detected to be on the book (Onbook) 171, a virtual furniture object pops up on top of the cube. Using a rotating motion, the user can 'browse' through the catalogue as different virtual furniture items pop up on the cube while the cube is being rotated. When the cube is picked up (Offbook), the last virtual furniture item that seen on the cube is picked up 172. When the cube is detected to be on the board (Onboard), the user can add the furniture to the cube by lifting the cube off the board (Offboard) 173.To re-arrange furniture, the cube is placed on the board (Onboard) with the "right arrow" marker as the top face. When the cube is detected as placed on the board, the user can 'pick up' the furniture by moving the cube to the centre of the desired furniture.

Referring to Figure 18, when the furniture is being 'picked up' (Offboard), the furniture is rendered on top of the cube and an audio hint is sounded 180. The user then moves the cube on the board to a desired position. When the position is selected, the user simply lifts the cube off the board to drop it into that position 181.

Referring to Figure 19, to delete furniture, the cube is placed on the board (Onboard) with the "x" marker as the top face 190. When the cube is being detected to be on the board, the user can select the furniture by moving the cube to the centre of the desired furniture. When the furniture is successfully selected, the furniture is rendered on top of the cube and an audio hint is sounded 191. The user then lifts the cube off the board (Offboard) to delete the furniture 192.

When a furniture is being introduced or re-arranged, a problem to keep in mind is the physical constraints of the furniture. Similar to reality, furniture in an Augmented 23 Reality world cannot collide with or 'intersect' with another. Hence, users are not allowed to add furniture when it collides with another.

Referring to Figure 20, one way to solve the problem of furniture items colliding is to transpose the four bounding co-ordinates 200 and the centre of the furniture being added to the co-ordinates system of the furniture which is being collided with. The points ptO, pt1 , pt2, pt3, pt4 200 are transposed to the U-V axis of the furniture on board. The U-V co-ordinates of these five points are then checked against the x- length and y-breadth of the furniture on board 201.

U_N = cos0( -X₀) + sinθ(Y_N -Y₀) V_N = sin Θ(X_N - X_Q ) + cos Θ(Y_N - Y₀ ) where

Only if any of the U-V co-ordinates fulfil UN < x-length && VN < y-breadth will the audio effect sound. This indicates to the user that they are not allowed to drop the furniture item at the position and must move to another position before dropping the furniture item.

For furniture such as tables and shelves in which things can be stacked on top of them, a flag is provided in their furniture structure called stacked. This flag is set true when an object such as a plant, hi-fi unit or TV is detected for release on top of this object. This category of objects allows up to four objects placed on them. This type of furniture, for example, a plant, then stores the relative transformation matrix of the stacked object to the table or shelf in its structure in addition to the relative matrix to the centre of the board. When the camera has detected top face "left arrow" or "x" of the big cube, it goes into the mode of re-arranging and deleting objects collectively. Thus, if a table or shelf is to be picked, and if stacked flag is true, then, the objects on top of the table or shelf can be rendered according on the cube using the relative transformation matrix stored in its structure.

Referring to Figure 21 , interaction between virtual objects may be in response to the spatial relationship of the cubes. The distance between two cubes is used to define the interaction between a story character and other virtual objects in the story scene. In the first screenshot, the user moves the small cube with the Chinese princess virtual object towards the larger cube with the planets virtual object. The system 210 constantly measures the spatial relationship between the small and larger cube. When the spatial relationship is within certain parameters, for example, the distance between the small and larger cube is sufficiently close, a response from the virtual objects occurs. In this example, the virtual object associated with the larger cube changes from the planets to a rose as depicted in the second screenshot.

Although a regular shape has been described for the marker, the marker may be of an irregular shape. Although a marker with a border has been described, it is envisaged that in some embodiments, no border is necessary. For an irregularly shaped marker, patterns for the markers may: • distinguish themselves from the remainder of the surface. That is, to differentiate the ID of the marker; • have a high contrast edge to be easily separated from the background. Therefore, the colour is not necessarily restricted to only black and white; and • have at least four feature points which are used for tracking. When calculating a transformation matrix, at least four feature points are identified. However, instead of corners other feature points may be used such as large black dots.

The described irregular tracking method may complement square marker usage. When the corners or edges of a square are occluded, the irregular tracking provides temporary support.

Although the interactive system 210 has been programmed using Visual C++ 6.0 on the Microsoft Windows XP platform, other programming languages are possible and other platforms such as Linux and MacOS X may be used. Although a Dragonfly camera 211 has been described, web cameras with at least 640 x 480 pixel video resolution may be used.

Although the system 210 has been described in one embodiment as software, it is possible for all software functionality to be hard-wired into a circuit which is connected to the electrical circuitry of the camera. Hence it is envisaged that the image processing functions of the computer software be performed by a camera alone.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the scope or spirit of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects illustrative and not restrictive.

Claims

THE CLAIMS:

1. An interactive system for interacting with a device in a mixed reality environment, the system comprising: an object having at least two surfaces, each surface having a marker; an image capturing device to capture images of the object in a first scene; and computer software to track the position and orientation of the object in the first scene by identifying a marker; wherein the computer software in response to manipulation of the object causes the device to perform an associated operation.

2. The system according to claim 1 , wherein at least two surfaces of the object are tracked to identify a marker for tracking the position and orientation of the object.

3. The system according to claim 2, wherein the marker used for tracking the position and orientation of the object is identified on a surface with the highest tracking confidence.

4. The system according to claim 3, wherein the surface with the highest tracking confidence is determined according to the extent of occlusion of its marker.

5. The system according to claim 1, wherein the computer software retrieves multimedia content associated with an identified marker, and generates a second scene including the associated multimedia content superimposed over the first scene in a relative position to the identified marker, to provide a mixed reality experience to a user.

6. The system according to claim 1 , wherein the device includes a television, DVD player, lighting, or an air conditioner.

7. The system according to claim 6, wherein associated operations include power on or off, volume control, dimming level control or temperature control.

8. The system according to claim 1 , wherein the device is a computer.

9. The system according to claim 8, wherein the computer software causes other software applications on the computer to perform associated actions or tasks.

10. The system according to claim 9, wherein other software applications include an MP3 player, a Media Player for playing video clips or movies, a Photo Album to display digital photos or an Internet web browser.

11. The system according to claim 10, wherein associated actions for the Media Player application include playing, pausing, fast forwarding or rewinding a video clip.

12. The system according to claim 11 , wherein translational movement of the object left or right instructs the Media Player to rewind or fast forward.

13. The system according to claim 11 , wherein rotating the object clockwise or anti-clockwise instructs the Media Player to rewind or fast forward.

14. The system according to claim 5, whereby if there is more one object, a first object is used as an anchor for relative positioning of the associated multimedia content, and a second object is used to operate the device or software application.

15. The system according to claim 5, wherein the associated multimedia content includes a virtual representation of the device, or the user interface of the software application.

16. The system according to claim 1 , wherein the computer software relies on a state transition model to determine a response to a manipulation of the object, and to determine an associated operation for the device to perform.

17. The system according to claim 16, wherein each physical action and physical property of the object is virtually coupled to a virtual method and virtual attribute.

18. The system according to claim 1 , wherein the marker includes a discontinuous border that has a single gap.

19. The system according to claim 18 wherein the marker comprises an image within the border.

20. The system according to claim 19 wherein the image is a geometrical pattern.

21. The system according to claim 20, wherein the pattern is matched to an exemplar stored in a repository of exemplars.

22. The system according to claim 19, wherein the colour of the border produces a high contrast to the background colour of the marker, to enable the background to be separated by the computer software.

23. The system according to claim 20, wherein the computer software is able to identify a marker if the border is partially occluded and if the pattern within the border is not occluded.

24. The system according to claim 1 , further comprising a display device to display the second scene at the same time the second scene is generated.

25. The system according to claim 24, wherein the display device is a monitor, television screen or LCD.

26. The system according to claim 24, wherein the display device is a view finder of the image capture device or a projector to project images or video.

27. The system according to claim 24, wherein the video frame rate of the display device is in the range of twelve to thirty frames per second.

28. The system according to claim 1, wherein the image capture device is mounted above the display device.

29. The system according to claim 28, where the image capture device and display device face the user.

30. The system according to claim 29, wherein the object is manipulated between the user and the display device.

31. The system according to claim , wherein multimedia content includes two dimensional or three dimensional images, video or audio information.

32. The system according to claim 1 , wherein the at least two surfaces of the object are substantially planar.

33. The system according to claim 32, wherein the at least two surfaces are joined together.

34. The system according to claim 32, wherein the object is a cube or polyhedron.

35. The system according to claim 1, wherein the object is foldable.

36. The system according to claim 1 , wherein the computer software is installed on a desktop or mobile computing device such as a Personal Digital Assistant (PDA), mobile telephone or other mobile communications device with a built-in computer processor.

37. The system according to claim 1 , wherein the image capturing device is a camera.

38. The system according to claim 37, wherein the camera is a CCD or CMOS video camera.

39. The system according to claim 37, wherein the camera, computer software and display device is provided in a single integrated unit.

40. The system according to claim 37, wherein the camera, computer software and display device is located in remote locations.

41. The system according to claim 1 , wherein the associated multimedia content is superimposed over the first scene by rendering the associated multimedia content into the first scene, for every video frame to be displayed.

42. The system according to claim 1, wherein the position of the object is calculated in three dimensional space.

43. The system according to claim 42, wherein a positional relationship is estimated between the display device and the object.

44. The system according to claim 1 , wherein the captured image is thresholded.

45. The system according to claim 44, wherein contiguous dark areas are identified using a connected components algorithm.

46. The system according to claim 45, wherein a contour seeking technique is used to identify the outline of these dark areas.

47. The system according to claim 45 wherein contours that do not contain four corners are discarded.

48. The system according to claim 45, wherein contours that contain an area of the wrong size are discarded.

49. The system according to claim 45, wherein straight lines are fitted to each side of a square contour.

50. The system according to claim 49, wherein the intersections of the straight lines are used as estimates of corner positions.

51. The system according to claim 50, wherein a projective transformation is used to warp the region described by the corner positions to a standard shape.

52. The system according to claim 51 , wherein the standard shape is cross- correlated with stored exemplars of markers to identify the marker and determine the orientation of the object.

53. The system according to claim 51, wherein the corner positions are used to identify a unique Euclidean transformation matrix relating to the position of the display device to the position of the marker.

54. The system according to claim 1 , wherein the marker is unoccluded to identify the marker.

55. The system according to claim 1 , wherein the marker is a predetermined shape.

56. The system according to claim 55, wherein at least a portion of the shape is recognised by the computer software to identify the marker.

57. The system according to claim 56, the computer software determines the complete predetermined shape of the marker using the recognised portion of the shape.

58. The system according to claim 57, wherein the predetermined shape is a square.

59. The system according to claim 58, wherein the computer software determines that the shape is a square if one corner of the square is occluded.

60. The system according to claim 1 , further comprising at least two objects, wherein the spatial relationship between the at least two objects is determined to cause a predetermined response from the multimedia content associated with the identified markers.

61. The system according to claim 60, wherein the spatial relationship is selected from the group consisting of: distance, stacking and occlusion between the objects.

62. The system according to claim 60, wherein the predetermined response is selected from the group consisting of: interaction between the associated multimedia content, animation of at least one associated multimedia content and playback of an audio recording for at least one associated multimedia content.

63. An interactive system for interacting with a device in a mixed reality environment, the system comprising: at least two objects, each object having at least two surfaces, each surface having a marker; an image capturing device to capture images of the objects in a first scene; and computer software to track the position and orientation of the objects in the first scene by identifying a marker on each object; wherein the computer software in response to manipulation of the objects and their arrangement relative to each other causes the device to perform an associated operation.

64. The system according to claim 63, wherein at least two surfaces of the object are tracked to identify a marker for tracking the position and orientation of the object.

65. The system according to claim 64, wherein the marker used for tracking the position and orientation of the object is identified on a surface with the highest tracking confidence.

66. The system according to claim 65, wherein the surface with the highest tracking confidence is determined according to the extent of occlusion of its marker.

67. The system according to claim 64, wherein the computer software retrieves multimedia content associated with an identified marker, and generates a second scene including the associated multimedia content superimposed over the first scene in a relative position to the identified marker, to provide a mixed reality experience to a user.

68. A method for interacting with a device in a mixed reality environment, the method comprising: capturing images of an object having at least two surfaces, each surface having a marker; and tracking the position and orientation of the object by identifying a marker; wherein in response to manipulation of the object, the device is made to perform an associated operation.

69. The method according to claim 68, wherein at least two surfaces of the object are tracked to identify a marker for tracking the position and orientation of the object.

70. The method according to claim 69, wherein the marker used for tracking the position and orientation of the object is identified on a surface with the highest tracking confidence.

71. The method according to claim 70, wherein the surface with the highest tracking confidence is determined according to the extent of occlusion of its marker.

72. The method according to claim 69, wherein the computer software retrieves multimedia content associated with an identified marker, and generates a second scene including the associated multimedia content superimposed over the first scene in a relative position to the identified marker, to provide a mixed reality experience to a user.

73. A method for interacting with a device in a mixed reality environment, the method comprising: capturing images of at least two objects, each object having at least two surfaces, each surface having a marker; and tracking the position and orientation of the objects by identifying a marker on each object; wherein in response to manipulation of the objects and their arrangement relative to each other, the device is made to perform an associated operation.

74. The method according to claim 73, wherein at least two surfaces of the object are tracked to identify a marker for tracking the position and orientation of the object.

75. The method according to claim 74, wherein the marker used for tracking the position and orientation of the object is identified on a surface with the highest tracking confidence.

76. The method according to claim 75, wherein the surface with the highest tracking confidence is determined according to the extent of occlusion of its marker.

77. The method according to claim 73, wherein the computer software retrieves multimedia content associated with an identified marker, and generates a second scene including the associated multimedia content superimposed over the first scene in a relative position to the identified marker, to provide a mixed reality experience to a user.

78. A software application for interacting with a device in a mixed reality environment, the application comprising: an image processing module to receive captured images of an object in a first scene from an image capturing device; and a tracking module to track the position and orientation of the object in the first scene by tracking at least two surfaces of the object where each surface has a marker, and identifying at least one marker; wherein the software application in response to manipulation of the object causes the device to perform an associated operation.

79. An image capturing device for interacting with a second device in a mixed reality environment, the device comprising: an image capture module to capture images of an object in a first scene; and a tracking module to track the position and orientation of the object in the first scene by tracking at least two surfaces of the object where each surface has a marker, and identifying at least one marker; wherein in response to manipulation of the object, the second device is made to perform an associated operation.

80. A computer program product comprised of a computer-readable medium for carrying computer-executable instructions for: receiving captured images of an object in a first scene from an image capturing device; and tracking the position and orientation of the object in the first scene by tracking at least two surfaces of the object where each surface has a marker, and identifying at least one marker; 0/22 wherein in response to manipulation of the object, a device is made to perform an associated operation.