GB2480173A - A data structure for representing an animated model of a head/face wherein hair overlies a flat peripheral region of a partial 3D map - Google Patents

Abstract

A data structure for representing an animated model of a head/face which may be rendered as an image on a display screen comprising: a map having a partial three dimensional surface defined as a mesh of interconnected nodes, the surface lying generally in a single plane in a peripheral region and being raised out of that plane in a central region corresponding to the topology of a face; a two dimensional image of a head/face which is viewed from the front which includes a face, surrounding hair and neck portion, and can be conformed to the surface of the map to provide a partial 3 dimensional model of the face in which the face has contours corresponding to the faces features and in which the hair overlies the flat peripheral region; at least one user defined co-ordinate corresponding to the location of a part of a facial feature in the model; and at least one facial feature which is located on the map in a position defined by the user defined co-ordinate.

Description

DATA STRUCTURE AND METHOD FOR PRESENTING MESSAGE

This invention relates to communication networks, and to a device for use in sending and receiving messages across a communication network. It also relates to a novel method of presenting a message to a user and to a data structure encoding information which can be used to present an image of a face associated with a sender of a message to a reader of the message.

At present, there are many different communication networks that are widely used to enable personal communication over long distances.

Traditionally the only form of communication was to send a letter or use a telephone but the latest trend has seen advances in instant or near instant written communication. Examples of such forms of communication are emails and text messages (or more correctly SMS or MMS messages).

In the case of emails and text messages a sender types a message into a sender device, such as a mobile phone or personal computer. The message is then sent across an electronic network to a receiver device. A user can then pick up the sent message and display the text on a display screen associated with the device.

Whilst these have proven very popular, especially with younger users, it has for some time been felt that these messages can be misinterpreted because they lack any way of expressing the emotions of the sender. They are also somewhat impersonal and difficult to read by users who have a visual impairment.

One partial solution to this problem has been to develop a system of symbols, known as emoticons, which can be included in a typed message.

These symbols represent expressions and help a reader determine the emotions intended to be expressed by a sender. For example, a "smiley" face can be inserted to show that the sender is happy.

It is an object of at least one aspect of the present invention to at least partially ameliorate the problem of including expression or other forms of personalisation in a typed message sent across a communication network such as a text message, email or instant message.

According to a first aspect the invention provides a data structure for representing an animated model of a head/face which may be rendered as an image on a display screen comprising: a map having a partial three dimensional surface defined as a mesh of interconnected nodes, each node being located on the surface and groups of nodes defining polygons which in turn define the contours of the surface, the surface lying generally in a single plane in a peripheral region and being raised out of that plane in a central region corresponding to the topology of a face; a two dimensional image of a head/face which is viewed from the front which includes a face, surrounding hair and a neck portion, and can be conformed to the surface of the map to provide a partial 3 dimensional model of the face in which the face has contours corresponding to the faces features and in which the hair overlies the flat peripheral region; at least one user defined co-ordinate corresponding to the location of a part of a facial feature in the model; and at least one facial feature which is located on the map in a position defined by the user defined co-ordinate.

This data structure provides an efficient model or representation of a face which can be used, inter alia, with the preceding aspects of the invention.

By laying an image on a partial 3D map, which is flat except for the region occupied by facial features where is raised out of the plane it gives a life like appearance when the model is rendered on a display, yet it requires little data to construct the 3D image when compared with a full 3D representation. Identifying the location of facial features allows animated features such as a mouth or eyes to be added during rendering to add realism, without needing to represent the feature in the data structure.

Having the flat region around the face makes it possible to display hair or other features beyond the face in a simple yet realistic way.

According to a second aspect the invention provides a method of producing an animated partial 3D model of a head for display on a display screen comprising the steps of: Selecting a map having a partial three dimensional surface defined as a mesh of interconnected nodes, each node being located on the surface and groups of nodes defining polygons which in turn define the contours of the surface, the surface lying generally in a single plane in a peripheral region and being raised out of that plane in a central region corresponding to the topology of a face; selecting a two dimensional image of a head/face which is viewed from the front which includes a face, surrounding hair and a neck portion fitting the image to the surface of the map to provide a partial 3 dimensional model of the face in which the face has contours corresponding to the face's features and in which the hair overlies the flat peripheral region; selecting from a data structure at least one user defined co-ordinate corresponding to the location of a part of a facial feature in the model; and selecting at least one facial feature; and locating the feature on the map in a position defined by the user defined co-ordinate.

The data structure and method of the invention may be used in a communication system comprising: a sending device a receiving device and a network which connects the sending device to the receiving device; in which the sending device comprises: at least one user operable input for entering a sequence of textual characters as a message and transmission means for sending the message across the network; in which the receiving device comprises: a memory which stores a plurality of head images, each one being associated with a different sending device and comprising an image of a head viewed from the front including a face, hair and neck portions; receiver means for receiving the message comprising the sequence of textual characters; text to speech converting means for converting the text characters of the message into an audio message corresponding to the sequence of text characters; animating means for generating an animated partial 3D image of a head from the head image stored in the memory which is associated with the sender of the message; the animating means animating at least one facial feature of the head, the animation corresponding the movements made by the head when reading the message; display means for displaying the animated partial 3D head; and loudspeaker means for outputting the audio message in synchronisation with the displayed head, and in which the memory includes a 3D mesh which is defined by a set of interconnected nodes which generally all lie in one plane around a peripheral region of the mesh and nodes within the periphery which are raised above the plane to correspond to facial features, the animating means generating the partial 3D head image by overlaying a head image onto the mesh with facial features of the facial region of the head image aligned with the raised facial features of the mesh and the hair portions overlaying the flat peripheral region of the mesh.

In general, much of the meaning and recognition conveyed in communication between humans is carried by the facial expression and the familiarity of one person with the facial appearance of the other. The representation of facial features in this invention is capable of great accuracy because it can be based on a digital photograph of the user or sender, which is already a good likeness and makes the animated partial 3D image described herein appear to the receiver as a realistic and recognisable representation of the sender of a message.

By performing the conversion of the text message into an animated and spoken message at the receive device there is no additional burden on the network when compared with the transmission and display of a text only message. Additionally, rendering the animated head from images prestored in the memory of the receive device removes the need to send the image with the transmitted message. Additionally, using a partial 3D rather than full 3D rendering reduces further the computational burden.

By partial 3D we mean that the animated displayed head is not a full 3D representation of a head. In a sense it may comprise a 2D image (the head image may be a 2D image such as a picture taken with a camera) which is deformed in places to give it some depth in the Z plane so that facial features protrude out of the 2D plane. Other parts of the image remain 2D. This partial 3D image is a 2D image which is distorted and made to appear 3D. The displayed image may be tilted slightly from left to right by simply changing the orientation of the base plane which can correspond with the periphery of the image. Because the facial features are given depth in the Z-plane, when tilted the image viewed on the display will appear to be truly 3D. The mesh may also be rotated in three planes to make the head appear to tilt to one side or slightly up or down in a nodding motion, or turn from side to side.

This mesh may therefore replicate the 3D topology that would result if a head was pressed into the back of a sheet of elastic material that is stretched taut across a frame. The material will be pushed forward by features of the face such as the nose and eyebrows and lips, yet remain in the same plane outside of the facial region.

This 3D mesh which is raised in regions of facial features but is flat outside of those regions allows for a head image which includes hair or other features beyond the outline of the face to be simply mapped onto the mesh. This is much simpler than producing a full 3D model and gives excellent results in terms of the realism achieved. With a 3D model, realistic presentation of hair is impossible to achieve. The raised features of the face allow the head image to be rotated slightly in three planes during animation and give the appearance of being truly 3D even though it is only partial 3D.

The mesh may have a generally rectangular outline to suit the rectangular outline of a typical rectangular display screen. This allows the animated image to be expanded to fill a display screen if desired.

Only a single mesh is stored in the memory of the receive device for use in rendering the animated head from any sender whose head image is stored in the receiver device. This reduces the amount of memory needed to provide an animated head image compared with storing many meshes, perhaps even one per head image. Of course, more than one mesh could be stored if desired.

The mesh may be modelled with a plurality of links connected to the nodes which mimic the attachment of a face to the bones of a skull, movement of the "bones" causing movement of the nodes relative to one another in the mesh to create animation.

A separate animation of the mesh may be stored for animating each speech phoneme.

A stored head image may comprise a photograph, such as a digital photograph, or other 2D image of a head (photographic or stylised) when viewed from the front. It may typically be a photograph of the senders head. The image may be sized such that the face is a set size that will match the size of the face in the 3D mesh. This can be achieved from any photo by cropping or zooming the image as required. For maximum realism the image should include a region around the face showing the hair and neck portions that will lie on the flat part of the 3D mesh.

The photograph may therefore be edited before use, perhaps to enhance or disguise a feature of the senders face. Alternatively it may be any photograph of a persons head/face such as an actor/actress or singer or other celebrity. This could be captured by a digital camera, or using a digital scanner. A sender can then choose to associate themselves with this head image.

Associated with each head image in the memory may be one or more co-ordinates which define the location on the mesh of a facial feature that is to be animated. This may, obviously, be the location of a mouth.

Co-ordinates of other features that may be animated may also be stored.

This may include eyes and eyebrows.

The memory may store one or more facial features that may be animated such as a mouth, eyes, eyebrows etc. Where more than one version of each feature is provided, for example two or more different eye socket shapes, a parameter may be associated with each head image to indicate which of the features is to be used in the animation.

Additionally the head image may be associated with an identifier which indicates the identity of a sender of a message associated with that face image.

A head image, coordinates and identifier may be grouped as a single data structure. This can then be easily transmitted from one device to another whenever the devices are connected on the network for the first time.

Typically the data will be transmitted via a server or other intermediary.

The transmitting device may transmit the identifier along with the message or as part of the message. In its simplest form, the identifier may comprise the unique network address (IP address, phone number etc) of the transmitter device. This identifier can then be matched to the correct head image at the receiver device.

The head image may be stored on the receive device prior to receipt of a message as part of an initial set-up process when it is first intended to receive a message from a new user. This process could be initiated by the receive device requesting a head image which is then sent by the transmitter. Alternatively it may be initiated by the person who wants to send a message to a receiver device for the first time.

Importantly the transmission of the head image does not happen again after initial set-up. This again means that no additional data needs to be sent with the text message. Of course, if the head image has changed it could be re-sent in its changed form if required, but if it does not change it need only ever be sent the one time during initial set up.

The converting means may include a dictionary which is stored in a memory of the receiving device which lists phonemes for different sequences of textual characters.

Where a dictionary is provided, it may include comparison means for comparing text in the message to words or sounds in the dictionary to construct the audio message.

The dictionary may also store, for one or more sounds (preferably for each and every sound) that will make up the audio message, an animation of a facial feature that corresponds to that sound which will be displayed by the animation means. This may comprise an animated mouth but may also include a pair of animated eyes or other feature such as eyebrows.

An alternative to the dictionary, which is more preferred, is to use a rule based text to speech conversion schema. This may be implemented by providing in the memory a set of rules which define the phoneme that is to be used for a given combination or sequence of textual characters. The memory may also include a set of exceptions which indicate sequences of textual characters that do not conform to the rules.

The use of a set of rules results in a more compact implementation of a text to speech converter compared with a full dictionary based system.

This is significant where the receiver device comprises a mobile device such as a telephone which has limited available memory compared with a larger desktop computer device. Rules may be provided for more than one language, and it is envisaged that almost any language can be converted from text to speech using a system according to the present invention provided sufficient rules are defined.

The audio message may comprise any audio format which is known in the art and which can be converted to an analogue audio signal by the receiving device. It may comprise a file in the.wav format for example.

The network may comprise a cellular telephone network and the sending device and receiving device may comprise a cellular telephone. It may comprise a fixed telephone network with fixed telephones for the sending and receiving devices. The messages may comprise text messages in the SMS or MMS format or similar.

More preferably the transmitted message may be sent in one of the standard instant messaging data formats such as XMPP and in particular Jabber. This is preferred as the transmission is faster than other mobile protocols such as SMS or MMS and cheaper.

Alternatively the network may comprise an internet or other form of communications network and the devices may comprise any devices which can send data across the internet such as PCIG, PDAs, laptops, tablet PCs, smart phones etc. The transmission means will vary dependent on which network the device is to be used with. For example, it may comprise an antenna for a GSM phone network, or an antenna of a wi-fi network or a data port for connecting to an internet.

It is beneficial to the sender of the message to know whether the intended recipient is present at his receiving device or absent. Where the sending device is in communication with a messaging server and may be equipped to indicate whether the recipient is on line, not willing to be disturbed, or absent. This enables the sender to choose whether he will be able to conduct a two-way message conversation with the intended recipient or whether he should simply send a one-way message. The server may include a facility to store messages whose intended recipient is absent, and subsequently forward the message when the recipient returns.

The display means may comprise a liquid crystal display which may be monochrome or colour. It should have a refresh rate sufficient to enable the face to be smoothly animated, e.g. greater than 12 frames per second.

The loudspeaker means may comprise a small speaker built in to the device or perhaps a detachable headphone, connected by a hard wire or a wireless link to the device.

It will be understood that all the key features such as display and loudspeaker and receiver and transmitter means can be found in a device such as a mobile phone. Hence one device can act as both a sender and a receiver device.

It is important that the message indicates the identity of the sender so that the receiver device can select the face to display that corresponds to the user of the sending device. In practice, this may comprise the telephone number of the sender device (for a telecommunication network) or an email address for an internet.

Typically many hundreds or thousands of sending and receiving devices may be connected to a network. In this case, a receiving device may store a database of different faces to display, each one corresponding to a different sender of a message.

By providing the animated head/faces, the presentation of a message is greatly enhanced. With the audio, this makes the device suitable for a new set of users such as those with impaired vision or with reading difficulties. It also makes the experience of reading a message more personalised as the identity of the user can be seen in the image.

The receiver device may also include an image generation means which displays a plurality of the head images on the display at the same time, with only one of them being animated by the animating means at any time. The generation means may cause the head images to be displayed such that the animated head image is displayed in a position such that it appears to be in front of the other head images. It may move the images around whenever one head image is to be animated to move that image to the front.

The generation means may display the head images in a circle which move around like a carousel as a head image is needed at the front for animation.

By displaying a plurality of heads at the same time, the user of the receiver device can easily identify the possible people they can communicate with across the network.

The sender device may also include a similar image generation means which displays head images on the screen at the same time. The user may manipulate the device to move a head image to the front to indicate that a message is to be sent to the receiver device associated with that head image.

The memory of the sender device (or receiver device) may include a group label associated with each head image, and the image generation means may display at the same time all head images which carry the same group label.

More than one group label may be associated with each head image, and the user may operate the device to select which group is to be displayed.

For example, a "work" group label and a "friends" group label may be provided.

The data structure and method of the invention may also be used to provide a communication device adapted to send and receive messages across a network comprising: a memory which stores a plurality of head images, each one being associated with a different sending device and comprising an image of a head viewed from the front including a face, hair and neck portions and in which the memory includes a 3D mesh which is defined by a set of interconnected nodes which generally all lie in one plane around the peripheral region of the mesh and nodes within the periphery which are raised above the plane to correspond to facial features, receiver means for receiving the message comprising the sequence of textual characters; text to speech converting means for converting the text characters of the message into an audio message corresponding to the sequence of text characters; animating means for generating an animated partial 3D image of a head from the head image stored in the memory which is associated with the sender of the message; the animating means generating the partial 3D head image by overlaying a head image onto the mesh with facial features of the facial region of the head image aligned with the raised facial features of the mesh and the hair and neck portions overlaying the flat peripheral region of the mesh, the animating means animating at least one facial feature of the head, the animation corresponding to movements made by the head when reading the message; display means for displaying the animated partial 3D head; and loudspeaker means for outputting the audio message in synchronization with the displayed head.

The device may include any of the optional features of the receiver device described in relation to the first aspect of the invention.

The device may include an additional dictionary of tags comprising symbols or sequences of symbols (textual or otherwise) which correspond to emotions. These are sometimes known as emoticons in the art. An example is the symbol © to show happiness and 0 to show sadness.

The device may be adapted, when identifying such a symbol, to cause the animated image of the face to express that emotion. For example, if an emoticon indicating that the sender is expressing happiness is identified it may cause the animated face to smile.

The dictionary may include a choice of different facial features, such as mouths, associated with each sound. Which one to use may be indicated by an identifier associated with the face that is to be displayed.

The device may include a speaker through which the speech can be reproduced. Alternatively, it may include an output port through which an audio signal can be passed to a speaker. An example of the later would be a headphone jack socket.

There will now be described, by way of example only, one embodiment of the present invention with reference to the accompanying drawings of which: Figure 1 is an overview of a communication network and connected devices in accordance with an embodiment of the present invention; Figure 2 is a schematic representation of a mobile telephone device in accordance with at least one aspect of the invention; Figure 3 illustrates a typical text message as displayed on the display of the device of Figure 2 during text entry.

Figure 4 is a flow diagram setting out the steps performed in creating a data structure representing a head image to be sent across a network, Figure 5(a) is an illustration of a 2D image of a face/head to be rendered; Figure 5(b) is a representation of a typical map used in constructing a rendered head/face image; Figure 6 shows the image conformed to the map as it is pushed through from the back of the image; Figure 7 illustrates the step of locating the corner of facial features in the mapped image; Figure 8 illustrates the complete data structure required to define a head image; Figure 9 is a flowchart illustrating the steps involved in presenting a message on a receiving device; Figure 10 is an overview of an alternative communication system which is in accordance with an aspect of the present invention; and Figure 11 illustrates the display of a set of head images at the same time in the form of a carousel.

As shown in Figure 1, a pair of processing devices 10,20 are connected across a network 30. The network comprises a cellular telephone network which can carry both audio and data messages between devices connected to the network.

For clarity, in the remainder of this description one device on the network will be referred to as a sending device 10, and another as a receiving device 20. The sending device 10 enables a user to send a message across the network 30. The receiving device 20 enables a user to receive a message sent across the network 30. In practice a single device can perform the functions of both a sender device and a receiver device 10,20.

A representative sending device 10 is shown schematically in Figure 2 of the accompanying drawing. It comprises a keypad 12 for entering commands and phone numbers and a display 14 such as an LCD for displaying data. It also includes a first area of non volatile memory 16 in which are stored program instructions, preferably made from flash memory, and can either be located inside a Subscriber Identity Module (SIM) card of the device 10 or located outside the SIM card as dedicated memory of the device 10. A processor 18 controls the operation of the device 10 in accordance with the instructions stored in the memory.

The memory 16 will also contain one or more messages which have been received from other devices, and one or more messages which are to be sent or have been sent from the device. These are typically arranged into folders, a so called "Inbox" and "Sent Items" folders. The user can select to view the contents of either folder using the keyboard and then to select a message from within the folder to display.

The messages in this example comprise instant message service messages in XMPP format, but could alternatively comprise messages in other networking protocols. An example of a message 40 which may be sent and reproduced is shown in Figure 3 of the accompanying drawings as displayed on a typical screen of the device of Figure 2 during text entry.

The message is entered by first selecting "new message" from the devices onscreen menu (not shown) and then entering each character through the keyboard 12. The keyboard comprises a reduced keyboard set having only 9 text keys and 3 function keys. Each text key carries several characters, and they can be selected using either a multi-stroke or two stroke entry strategy. Such strategies for entry of text on reduced keyboards are well known in the art.

Such processing devices are widely known in the art, and it is also known to provide a facility to load additional programs into the memory 16.

These can then be called up by a user through the keyboard 12, and when running on the processor 18 of the device cause it to perform additional functions. As shown in Figure 2 the device includes in its memory two programs. One is a communication program 22 which enables the device to send and transmit audio or data messages across the network 30. The other is a novel program 24 called an Amego program which augments the presentation of text messages so as to enhance the users interaction with the device. In practice, many other programs may be stored such as a calendar or calculator program.

In this example, the program 24 which is stored in the memory 16 enables the phone to communicate with other devices across the network 30 in a new manner. Specifically, it enables the user to send to a remote device a partial-3D representation of their head (or another real or imaginary persons head or a modified form of their head) which can be displayed on the display of the device such that it gives the appearance of "reading-out" the senders messages sent across the network. The program can be written in any known programming language which is supported by the sender device, and the invention should in no way be construed as limited to any particular programming language. It may, for example, be written in Java.

The program when executed on the processor of the device causes the device to perform several functions: (1) It enables a user to create a head image or at least it stores the definition of a face/head image in the memory of the device; (2) It sends a data structure representative of a face/head image to a user either in response to a request from the user, or with each message, or on initial contact with a new user; (3) It renders on the display on the device of an animated image of a face/head as instructed by a remote device and defined by a data structure; (4) It reads incoming messages and converts them into speech which can be read out to the user of the device, and animates a displayed face/head image to match the sender of the message.

The Four key features of the program will now be explained in turn: (1) Creating of a head image.

The device when running the program enables a user to create a data structure defining an image of a face/head to be sent to a remote device.

It stores this data structure in memory in such a manner that it can be accessed and then be readily transmitted across the network to another device whilst requiring relatively little bandwidth.

To create the rendered face/head image the program performs the functional steps set out in Figure 4 of the accompanying drawings. In a first step 41, the user prompts the device to start the creation of a new head. The device then prompts 42 the user to provide an image, in two dimensions, of a head viewed from the front so as to show a complete face. This image may typically be a photograph captured by a digital camera or scanned from a printed image. A sample image is shown in Figure 5(a) of the accompanying drawings.

In a next step 43, the image is mapped onto a map of a three dimensional surface stored in the memory of the device. The map is defined as a mesh of interconnected nodes that define a surface. The map is generally rectangular. The location of the nodes can be defined in relation to their position (X-Y coordinates) relative to a base plane (Z co-ordinate of zero) and the plurality of nodes may be located at different heights relative to the base plane and spaced apart across a region of the base plane corresponding to a defined head profile. A sample mapped surface is shown in Figure 5(b) of the accompanying drawings as a series of polygons.

The map is not a full 3D representation of a head, but a partial 3D representation. It is a partially flattened representation of how a model of a head would appear when viewed from the front or slightly to each side.

As such it defines a nose and eyebrows which are partially flattened compared with real features. This has been found to be perfectly acceptable for a model of a head which is to be shown from the front only, the partially flattened features providing some perspective and shading to the finished head image.

The nodes of the mesh generally lie in the base plane all round the periphery of the map. This region is therefore flat. The inner boundary of this flat region correspond generally to the outline of a head and face.

Within the head/face region, the nodes rise up out of the base plane to define facial features such as a nose and eyes and mouth and cheekbones, forehead and the like.

The step of conforming the image to the 3D surface can be thought of as r.. . analogous to printing the image on a sheet of perfectly elastic material suspended in a frame, and pressing the map into the back of the image so that the map pushes the image out of its original flat plane to conform to o the topological features of the map. The image is pushed until the base plane of the map coincides with the plane of the image. This is shown in Figure 5(b) of the accompanying drawings. To help this process, the user is shown the map behind the image prior to conforming to allow the user to align the map and image on the display. The user can also scale the image to fit the map if required. The map is then "pushed" into the back of the image to make it a partial three dimensional model.

The end result is the facial region of the image is given contours whilst the region around it, which may contain hair for example, is flat. This partial 3D model can, when displayed, be tilted slightly to the side and because it has depth in the Z-plane will appear to be a real 3D head rotating in three planes. This can be seen in Figure 6 of the accompanying drawings.

In the next stage 44, the partial 3D image is displayed on the display and the user is prompted to indicate 45 the location co-ordinates of opposing corners of the mouth in the displayed image. This can be performed by the device showing cursors on the screen which the user can position using the keypad. The user is then asked to indicate when they are happy with the position of the cursors. Once this is complete, the co-ordinates of the corners are stored in the memory. The user is then prompted to select 46 from a choice of different features to be provided at the coordinates, e.g. different shaped eye sockets or colours of eyes. This is shown in Figure 7 of the accompanying drawings.

In the next step 47, the user is prompted to indicate the location of the centre of both eyes in the displayed mapped image.

Finally, the original 2D image and the coordinates of the mouth and eyes are stored as a data structure which represents the complete head image.

This may comprise a single electronic file. It is to be noted that provided knowledge of the original topological map is known, this is all that is needed to recreate the head image on any device. The data is simply fitted to the standard map.

Optionally, the user is also presented with a selection of different facial feature such as eye socket shapes from which to select. These comprise pre-programmed animated features, and the selection allows the user to choose the preferred eye colour or shape and so on. The identity of the chosen feature is then stored in the data structure.

The content of the data structure is illustrated in Figure 8 of the accompanying drawings.

Of course, it is possible for this creation stage to be executed on a device other than a mobile device, such as a personal computer. This may have advantages in terms of ease of use, as a personal computer will often have a more comprehensive user interface than a mobile device. For example, a PC may include a mouse which simplifies the task of the user indicating the location of the mouth and eyes.

(2) Sending the head image to a remote device.

Once a data structure representing a head image has been created it may be stored on the sender device and transmitted to a central network server where it is stored in a database along with other details of the sender, including unique means of identification. When another user is in correspondence with the sender for the first time, this user will receive the sender's head image data structure and have it stored on his own device, so that in future correspondence the data structure does not need to be sent again, thus reducing the volume of data transmitted when exchanging subsequent messages. The unique means of identification may be a unique number, IP address or MAC address and will be associated exclusively with the head image data structure in a database so that messages can be correctly routed. In the event that users make changes to their head image the sending device can send the amended image to the database, which can then forward the amended image to all those receiving devices that it has recorded to have previously been in contact with that sender.

The head image is sent to the remote device by sending the 2D image and the co-ordinates of the corners of the salient features, e.g. mouth and eyes. This is advantageous as it can be compressed into a small file size compared with sending a full 3D image.

At the remote device, the data structure is stored in a memory indexed by the identity of the sender of the data structure. This is important as it enables the remote device to choose the correct data structure when a message is later received.

(3) Rendering an image and reading a message An important aspect of the program is its ability to render a face/head image on the display of the device and to animate this in time with a spoken form of a text message such that the head appears to read the message. This is really a function of a receiver device and is invoked whenever a user selects a message that is stored in the memory of the device and prompts the device to display the message.

The steps of presenting the spoken and animated message to a user are shown in Figure 9 of the accompanying drawings. In a first step 91, the identity of the sender of a message is determined. This is then checked 92 against a table of senders to see if it is one that is recognised. If it is in the table, then the corresponding head image is selected.

In the next stage 93, the message is analysed either letter by letter or word by word to identify phonemes in the message. The identification process may be achieved using a phonic dictionary stored in the memory of the device. This comprises a database of sound files, each of which corresponds to a phoneme. Using this database, an audio file is constructed that corresponds to the message that is to be read. However, in this embodiment it is converted using a rule based schema in which a set of rules stored in the memory is used to determine what sounds should be used for different sequences of textual characters.

In the next stage 94, the mouth is selected for the head image and is animated such that it appears to move in sync with the phonemes in the audio file. In practice, each phoneme will be associated with a sequence of mouth images or visemes which appear to show the mouth moving as if making the sound. The sequence of mouth movements is stored in a data structure.

In a final step 95, the head image is rendered on the display with the mouth image overlaid in sync with the playing of the audio file through the speaker of the device.

Because the map and the animations can be prestored in a memory of the device, all that needs to be received to add a new face/head image for a sender is a 2D image file and the location of the facial features. This represents far less data than receiving a complete 3D animation of the face from the sender, especially when faces for multiple senders may be displayed.

Modifications Various refinements to the invention are envisaged. In one refinement, an emotion dictionary may be provided in the memory of the device. This will include a set of predefined facial expressions identified by sequences of characters -tags -that may be typed in a message. These sequences are sometimes referred to as emoticons. For each different facial expression the facial features will be modified to show that expression.

The emotion dictionary and phoneme dictionaries may be combined.

Rather than a single mouth animation for each phoneme there may be several with each one corresponding to a different emotion. The correct one will be selected according to any emoticons inserted into the message.

For example, if a message starts with the tag/emoticon: ) then the mouth animations corresponding to a happy face may be used to animate the mouth. If it starts with the emoticon: ( then a different set of mouth expressions corresponding to a sad face may be used.

Additionally, the tone of the audio may be varied to correspond with the emotion.

In a still further refinement, the device may include an image generation means which causes a set of head images to be displayed on the screen at the same time. The images are displayed in the manner of a carousel.

This can be seen in Figure 11 of the accompanying drawings. Whenever a head image is to be animated the head images rotate around the carousel r. . . until that image is located at the front. To make it stand out further this one may be larger than the others or may be of a higher intensity.

o The device also permits the user to move the head images round the carousel to allow a head image to be selected. A message can then be entered and it will be sent by the device to the selected head image.

It will also be understood that whilst a preferred embodiment relates to mobile telephones the invention has much wider ranging application than that. In an alternative embodiment the messages comprise email and both the sender device and receiver device comprise personal computers (desktop computers, laptop computers, tablet PCs or PDAs) connected to one another across an internet. This is shown in Figure 10 of the accompanying drawings.

In this arrangement, rather than specifying a phone number as a destination for an SMS message, an email address is specified and the message takes the form of an animated partial 3D image embedded in an email message or as an attachment to an email message.

As before, the transmission of the 3D head may be achieved by embedding a file containing the information within the body of an email or by attaching the file to the email.

It is envisaged that the method of providing head images for use with the audible reproduction of a typed message will soon become a common worldwide standard. All that is required is to load the program onto a device which has a display and can receive messages. The program will include the standard topological map and a dictionary of phonics for use in text to audio conversion. Because only the 2D image and the co-ordinates of facial features need to be transmitted to send a face/head image to a remote device, a high bandwidth is not required to provide a partial 3D animation of the face/head.

In another alternative embodiment, an animated partial 3D image or plurality of partial 3D images may be embedded in a web page by means of a browser plug-in, in order to deliver a spoken message or other text or spoken information to the recipient, or to engage in a discussion between the senders of the images through the medium of the display devices.

In a further alternative embodiment, an animated partial 3D image or plurality of partial 3D images may be embedded in an otherwise independent software program, for example an email client or slideshow application or other program, in order to deliver a spoken message or other text or spoken information to the user or viewer through the medium of the display device. 2?

Claims (8)

1. CLAIMS1. A data structure for representing an animated model of a head/face which may be rendered as an image on a display screen comprising: a map having a partial three dimensional surface defined as a mesh of interconnected nodes, each node being located on the surface and groups of nodes defining polygons which in turn define the contours of the surface, the surface lying generally in a single plane in a peripheral region and being raised out of that plane in a central region corresponding to the topology of a face; a two dimensional image of a head/face which is viewed from the front which includes a face, surrounding hair and a neck portion, and can be conformed to the surface of the map to provide a partial 3 dimensional model of the face in which the face has contours corresponding to the faces features and in which the hair overlies the flat peripheral region; at least one user defined co-ordinate corresponding to the location of a part of a facial feature in the model; and at least one facial feature which is located on the map in a position defined by the user defined co-ordinate.
2. 2. A method of producing an animated partial 3D model of a head for display on a display screen comprising the steps of: Selecting a map having a partial three dimensional surface defined as a mesh of interconnected nodes, each node being located on the surface and groups of nodes defining polygons which in turn define the contours of the surface, the surface lying generally in a single plane in a peripheral region and being raised out of that plane in a central region corresponding to the topology of a face; selecting a two dimensional image of a head/face which is viewed from the front which includes a face, surrounding hair and a neck portion fitting the image to the surface of the map to provide a partial 3 dimensional model of the face in which the face has contours corresponding to the face's features and in which the hair overlies the flat peripheral region; selecting from a data structure at least one user defined co-ordinate corresponding to the location of a part of a facial feature in the model; and selecting at least one facial feature; and locating the feature on the map in a position defined by the user defined co-ordinate.
3. 3. The method of claim 2 which comprises the further steps of rendering the model defined by the data structure on a display.
4. 4. The method of claim 3 in which the step of rendering includes providing a rendered, animated, model of at least one facial feature such as mouth or eyes, and locating them on the displayed conformed image at the co-ordinates indicated in the data structure.
5. 5. The method of claim 4, which further comprises animating the model along with an audible reproduction of a typed message.
6. 6. The method of claim 3, 4 or S in which the step of rendering comprises embedding the animated model in a webpage.
7. 7. The method of any of claims 1 to 6 in which the step of rendering comprises rotating the animated model of the head slightly in three planes to give an illusion when presented on a display of the animated model being fully three dimensional.
8. 8. A data structure and method substantially as described herein with reference to and as illustrated in the accompanying drawings.