Abstract: An image processing apparatus (2) is disclosed in which input images are processed to generate texture map data for texture rendering a generated three-dimensional computer model of object(s) appearing in the images. In order to select the portions of the images utilized, confidence data is generated indicative of the extent portions of the surface of a model are visible in each of the images. The images are then combined utilizing this confidence data, where image data representative of different spatial frequencies are blended in different ways utilizing the confidence data.

Abstract: Creation of a computer representation of a three-dimensional object surface is described. The viewing cones for camera positions at which images of the object were taken are determined and the intersection of these viewing cones is used to define an initial three-dimensional space within which the object surface lies. This initial space is divided into voxels. Each non-occluded voxel is checked for photoconsistency by comparing the colors (or average colors) of the pixel patches in the images to which that voxel projects. Photo-inconsistent voxels are removed. A voxel may be determined to be photo-inconsistent if there exists no set of photoconsistent pixel regions or the pixel patches do not share a color value range. The pixel patch into which each voxel projects in a further image may be compared with the stored color for that voxel and any photo-inconsistent voxels removed. This process can then be repeated for further images.

Abstract: To generate texture data for a 3D computer model 150 of a real-life object 210, images 300–316 of the subject object are recorded from different viewing positions and directions. The image data is processed by an image processing apparatus 2 to register the viewing positions and directions of the images 300–316 based on the positions of features in the images. The 3D computer model 150 of the subject object is registered with the resulting registered set of images, and texture data is generated for the 3D computer model from the images.

Abstract: An apparatus (2) for matching features in images of objects taken from different viewpoints is provided comprising: an image buffer (60) for receiving image data; and output buffer (62) for outputting pairs of matched features and processing means (64–78) for processing received image data to determine matched pairs of features in images.

Abstract: In an image processing apparatus 20 an input sequence 130 of video images is processed to determine the different positions and orientations at which the images were recorded in an efficient and accurate manner. A subset of the input images are selected as keyframes to form a sequence 250 of keyframes. Respective triples of keyframes having different, non-overlapping positions in the sequence 250 are selected and processed to determine the relative positions and orientations at which the keyframes in each triple were recorded to form respective sets of keyframes. The positions and orientations of keyframes between the keyframes in each triple are then calculated to form expanded sets of keyframes 266, 276, 286. The sets are further expanded by calculating the positions and orientations of keyframes which lie between sets in the sequence 250. The sets are merged by calculating the relationship between the coordinate systems in which the positions and orientations of the keyframes in each set are defined.

Abstract: To generate a 3D computer model of a subject object 210, images 300-316, 380-384 of the subject object are recorded from different viewing positions and directions. The image data is processed to generate a plurality of sets of images, each set containing images having registered imaging positions and directions. A preliminary 3D computer model 390 is generated using the images from a first of the sets such that the position and orientation of the preliminary 3D computer model 390 is registered with the images in the set. The imaging positions and directions of the images in the first set are then registered with the imaging positions and directions of the images in each other respective set.

Abstract: In a computer processing apparatus 3002, a number of depth maps 3200-3270 of a subject object 3300 are processed to generate a 3D computer model of the subject object. The points in each depth map are connected to give a 2D mesh, and each 2D mesh is then projected into 3D space in dependence upon the depths of the points in the mesh, thereby giving a 3D mesh 3610. Side faces for each edge of the 3D mesh are added extending away from the depth map, thereby generating a respective polyhedron 3600 for each depth map. The 3D computer model of the subject object is generated by calculating the intersections of the polyhedra.

Abstract: A 3D computer model of an object is generated by calculating the intersections of polyhedra. Each polyhedron defines a volume of 3D space containing at least part of the object. The 3D points of intersection of the planar faces of the polyhedra are calculated and each point is labeled with the planar faces which meet thereat. The points are connected to form a polygon mesh using the labels to determine which points should be connected together. In calculating the points, a volume containing the object is subdivided into parts, each part is tested against the polyhedra and then discarded, subdivided further, or the point of intersection of planar faces within the volume part is calculated. A volume part is discarded if it is outside at least one polyhedron. The volume part is subdivided into further parts if it is intersected by more than a predetermined number of polyhedra faces.

Abstract: A texture map for texturing the polygon mesh of a 3D computer model during rendering is generated by defining a respective triangle within the texture map for each triangle in the polygon mesh to create a texture coordinate map, and allocating image data to each defined triangle. To generate the texture coordinate map, the triangles are defined so that the area of each triangle is dependent upon the content of texture data to be stored therein. More particularly, triangles required to store texture data with a relatively large amount of detail have a relatively large area and triangles which are required to store texture data with relatively little detail have a relatively small area. In this way, more area is allocated for the storage of detailed texture data, thereby reducing the amount of information which is lost from the texture data during the creation of a texture map.

Abstract: An image-based rendering method for processing depth map images of a scene recorded from different viewpoints, and generating a virtual image of the scene from an arbitrary viewpoint. To calculate the color value of a pixel in the virtual image, a pixel-viewing ray is defined from the focal point of the virtual camera through the pixel. The ray is projected into each depth map image, giving a projected ray. Pixels in a depth map image which are intersected by the projected ray are tested by defining a point (W1, W2, W3) along a line from the focal point of the depth map camera through the pixel at a distance corresponding to the depth of the pixel. A color value for the pixel in the virtual image is calculated by interpolating between the values of the pixels in the depth map image which produced the points (W2, W3) lying on opposite sides of the pixel viewing ray.

Abstract: In an image processing apparatus 2, images of a subject object 210 and data defining the positions and orientations at which the images were recorded are processed to generate a three-dimensional computer model of the subject object 210. As part of the processing, image data relating to the subject object 210 is segmented from other image data in each input image to define the silhouette of the subject object in each image, and the silhouettes are processed to generate the three-dimensional computer model.

Abstract: There is disclosed an apparatus and method for combining three or more texture maps 200-208 together to generate a single, combined texture map 220. The texture maps are repeatedly combined together in pairs, with each pair generating a larger texture map for subsequent combination with another texture map, until only one texture map remains containing all of the original texture maps.

Abstract: To generate a 3D computer model of a subject object 210, images 300-312 of the subject object are recorded from different viewing positions and directions. The image data is processed to calculate the positions and directions, and a 3D computer model of the surface shape of the subject object 210 is generated. A further image 330 of the subject object, which was not used in the processing to generate the 3D computer surface shape model but which shows a part of the subject object for which texture data is to be generated, such as the base of the subject object, is registered with the 3D computer surface shape model. The registration is carried out by displaying the further image to a user as a stationary image and providing a virtual camera for control by the user to view the 3D computer surface shape model. Images of the 3D computer surface shape model are generated and displayed overlaid on the stationary image in accordance with the virtual camera as it is moved by the user.

Abstract: In a 3D graphics processing apparatus 2, a virtual camera 100 is controlled to view a 3D computer model 120 in accordance with user instructions. The virtual camera 100 is constrained to move to different positions on a sphere around the 3D computer model in accordance with user instructions while having a viewing direction towards the centre of the sphere.

Abstract: In order to edit texture data for a 3D computer model, an image showing the 3D computer model with the texture data applied thereto is generated from a user-selected viewing direction and displayed to the user. Changes are made to the image data in accordance with user instructions. Corresponding changes are made to the texture data. To preserve texture data that the user did not wish to change and to retain the quality of the original texture data, processing is performed to amend the texture data which corresponds to only the image data that was changed and not texture data which corresponds to image data unchanged by the user. In addition, processing is performed to identify each polygon in the 3D computer model which is at an oblique angle to the viewing direction of the image displayed to the user. The identified polygons are then excluded from subsequent processing so that the texture data therefor is not changed.

Abstract: To generate texture data for a 3D computer model 150 of a real-life object 210, images 300-316 of the subject object are recorded from different viewing positions and directions. The image data is processed by an image processing apparatus 2 to register the viewing positions and directions of the images 300-316 based on the positions of features in the images. The 3D computer model 150 of the subject object is registered with the resulting registered set of images, and texture data is generated for the 3D computer model from the images.

Abstract: In a computer processing apparatus 2, a 3D computer model comprising a polygon mesh 500 representing the visual hull of an object 300 is generated by processing images of the object recorded at different positions and orientations to back-project the silhouette of the object in each image to give a respective cone which constrains the volume of 3D space occupied by the object. To remove concave and convex artefacts 510 in the polygon mesh 500, the polygon mesh is projected into each image to give a respective reference silhouette for each image. A change is made to at least one edge or vertex in the polygon mesh to give a refined polygon mesh, which is then projected into each image. The resulting silhouette in each image is tested against the corresponding reference silhouette.

Abstract: An image processing apparatus (2) is disclosed in which input images are processed to generate texture map data for texture rendering a generated three-dimensional computer model of object(s) appearing in the images. In order to select the portions of the images utilized, confidence data is generated indicative of the extent portions of the surface of a model are visible in each of the images. The images are then combined utilizing this confidence data, where image data representative of different spatial frequencies are blended in different ways utilizing the confidence data.

Abstract: A 3D computer model of an object is generated by calculating the intersections of polyhedra. Each polyhedron defines a volume of 3D space containing at least part of the object. The 3D points of intersection of the planar faces of the polyhedra are calculated and each point is labeled with the planar faces which meet thereat. The points are connected to form a polygon mesh using the labels to determine which points should be connected together. In calculating the points, a volume containing the object is subdivided into parts, each part is tested against the polyhedra and then discarded, subdivided further, or the point of intersection of planar faces within the volume part is calculated. A volume part is discarded if it is outside at least one polyhedron. The volume part is subdivided into further parts if it is intersected by more than a predetermined number of polyhedra faces.

Abstract: In a computer processing apparatus 3002, a number of depth maps 3200-3270 of a subject object 3300 are processed to generate a 3D computer model of the subject object. The points in each depth map are connected to give a 2D mesh, and each 2D mesh is then projected into 3D space in dependence upon the depths of the points in the mesh, thereby giving a 3D mesh 3610. Side faces for each edge of the 3D mesh are added extending away from the depth map, thereby generating a respective polyhedron 3600 for each depth map. The 3D computer model of the subject object is generated by calculating the intersections of the polyhedra.