Abstract: The invention relates to a method for acquiring a substantially complete depth map from a 3-D scene. Both depth values and derivates of depth values may be used to calculate a pixel dense depth map with the steps of acquiring partial depth map from said 3-D scene, acquiring derivates of depth information from said scene, and extending said partial depth map by adding non-relevant information to said partial depth map, creating a pixel dense full depth map being spatially consistent with both said partial depth map and said derivates of depth information.

Abstract: The present invention relates to a method, apparatus and computer program product for providing contour information related to images. An image obtaining unit obtains a set of interrelated images (step 26), an image segmenting unit segments said images, (step 28) and a contour determining unit (22) extracts at least two contours from the segmentation (step 30), selects interest points on the contours of each image (step 32), associates interest points with corresponding reconstructed points by means of three-dimensional reconstruction (step 34), projects reconstructed points into the images (step 36), and links reconstructed points not projected at a junction or their projections to each other in order to provide a first set of links (step 38), such that at least a reasonable part of a contour of an object can be determined based on the linked reconstructed points.

Abstract: The invention relates to a system (120) that makes recommendations on what to wear. Clothes (101–105) are tagged (106–110) so that taking them out of storage (100) can be detected. Using this information, a user profile is updated. This way, the system (120) learns from the user's behavior and is able to generate more accurate recommendations over time. The user profile can be stored on a smart card (320), so a user can use it in a shop to buy new clothes that fit with his current set (101–105). The user profile can also be used for targeted advertising.

Abstract: A method for matching digital images, including regularization of image features of a first digital image, composed of pixels, defining a finite set of candidate values, wherein a candidate value represents a candidate for a possible match between image features of the first image and a second image, establishing a matching penalty function for evaluation of the candidate values, evaluating the matching penalty function for every candidate value, selection of a candidate value based on the result of the evaluation of the matching penalty function, regularization of the first image by segmentation of the first image, comprising assigning at least part of the pixels of the first image to respective segments, determining a certainty parameter for at least part of the pixels of a segment, and establishing the matching penalty function to be at least partially based on the certainty parameter.

Abstract: In the method of motion estimation a tree (120) of segments (102–114) of an image (100) is generated by performing a hierarchical segmentation. The tree (120) of segments is analyzed to control the generation of a set (118) of candidate motion vectors of a segment (104). For the motion vectors of the set (118) match penalties are calculated. Finally, a particular motion vector (116) is selected from the set (118) of candidate motion vectors on the basis of match penalties. In the method of depth estimation depth data is calculated on the basis of a motion vector and the rules of parallax.

Abstract: A method of and system for providing a user profile, comprising user profile maintenance means for maintaining the user profile, visualization means for representing a content item visually as an icon and representing the user profile as a corresponding shape comprising at least one icon representing a content item, and consistency guarding means for guarding the consistency between the user profile and the corresponding shape. The consistency is guarded by changing the corresponding shape in dependence on a change in the user profile, and/or by changing the user profile in dependence on a manipulation of the corresponding shape.

Abstract: In block-based motion or depth estimation, a block is assigned a motion or depth value as a result of minimizing the matching error over a limited set of candidate values. The matching error for each element of the set is obtained by computing the luminosity differences between a block of a first image (10) and an area of a second image (11). It may occur that an object (12) is partially obstructed by another object (15), so that a pixel (14) in the block is not present in the corresponding area, because another pixel (16) overlaps it. The method and system according to the invention determine which pixels are not visible in the second image (11), and compute the matching error over only the visible pixels. An apparatus for adapting a video signal (40) uses the chosen candidate values to create an enhanced version of the video signal (40).

Abstract: A method for segmenting a digital image composed of pixels, including determining hard border fragments in the digital image by means of an edge detection process, determining respective sides of each of said hard border fragments, and forming segments by determining for each pixel a closest side and by uniquely assigning each pixel of said digital image to a single segment, corresponding with the determined closest side to the respective pixel.

Abstract: The invention relates to a method for providing an image to be displayed on a screen 10 such that a viewer in any current spatial position O in front of the screen 10 can watch the image with only a minimal perspective deformation. According to the invention this is achieved by estimating the current spatial position O of a viewer in relation to a fixed predetermined position Q representing a viewing point in front of the screen 10 from which the image could be watched without any perspective deformation and by providing the image by applying a variable perspective transformation to an originally generated image in response to said estimated current position O of the viewer, such that the viewer in said position O is enabled to watch the image without a perspective deformation.

Abstract: In block-based motion or depth estimation, a block is assigned a motion or depth value as a result of minimizing the matching error over a limited set (45) of candidate values (20, 21, 22, 23). When the chosen candidate value (23) is an extreme value of the set, it is impossible to predict how remote this value is from the real minimal matching error. To overcome this problem, the method and system for choosing an optimal candidate value choose a new candidate (30) beyond the extreme value (23). When this new candidate (30) then has the lowest matching error, additional new candidates are chosen until it no longer has. An apparatus for adapting a video signal (40) uses the chosen candidate values to create an enhanced version of the video signal (40).

Abstract: A computer graphics system (300) arranged to render polygons comprises: a frame memory (310) to store values of pixels; a scan conversion unit (304) for determining which of the pixels are inside polygons (100) and to detect Jags on the edges of the polygons; and a delineating unit (306) for delineating the polygons, being arranged to eliminate Jags (108-120) on the edges (102) of polygons (100) by mixing values of at least two of the pixels to achieve an appropriate value of a particular pixel to be displayed.

Abstract: In the method of motion estimation a tree (120) of segments (102-114) of an image (100) is generated by performing a hierarchical segmentation. The tree (120) of segments is analyzed to control the generation of a set (118) of candidate motion vectors of a segment (104). For the motion vectors of the set (118) match penalties are calculated. Finally, a particular motion vector (116) is selected from the set (118) of candidate motion vectors on the basis of match penalties. In the method of depth estimation depth data is calculated on the basis of a motion vector and the rules of parallax.

Abstract: A method of generating a three-dimensional representation (904) of at least one object (916) from multiple two-dimensional images (912) of the object makes use of an octree (902) of cells (903) to hold the three-dimensional representation (904), with each cell comprising vertices (906) and edges (910) connecting the vertices. The method is based on a process of splitting cells of the octree into smaller cells. A stop criterion for the process of splitting cells is based on inspecting which of the vertices of the cell are inside and which of the vertices are outside the object. Another stop criterion for the process of splitting a cell is based on inspecting whether the vertices of neighboring cells, are inside or outside the object.

Abstract: A method of and system (100) for providing a user profile, comprising user profile maintenance means (110) for maintaining the user profile, visualization means (120) for representing a content item visually as an icon (101) and representing the user profile as a corresponding shape (102) comprising at least one icon (101) representing a content item, and consistency guarding means (130) for guarding the consistency between the user profile and the corresponding shape (102). The consistency is guarded by changing the corresponding shape (102) in dependence on a change in the user profile, and/or by changing the user profile in dependence on a manipulation of the corresponding shape (102).

Abstract: The invention relates to a method for providing an image to be displayed on a screen 10 such that a viewer in any current spatial position O in front of the screen 10 can watch the image with only a minimal perspective deformation. According to the invention this is achieved by estimating the current spatial position O of a viewer in relation to a fixed predetermined position Q representing a viewing point in front of the screen 10 from which the image could be watched without any perspective deformation and by providing the image by applying a variable perspective transformation to an originally generated image in response to said estimated current position O of the viewer, such that the viewer in said position O is enabled to watch the image without a perspective deformation.

Abstract: A method for matching digital images, including regularization of image features of a first digital image, composed of pixels, defining a finite set of candidate values, wherein a candidate value represents a candidate for a possible match between image features of the first image and a second image, establishing a matching penalty function for evaluation of the candidate values, evaluating the matching penalty function for every candidate value, selection of a candidate value based on the result of the evaluation of the matching penalty function, regularization of the first image by segmentation of the first image, comprising assigning at least part of the pixels of the first image to respective segments, determining a certainty parameter for at least part of the pixels of a segment, and establishing the matching penalty function to be at least partially based on the certainty parameter.

Abstract: A method for segmenting a digital image composed of pixels, including determining hard border fragments in the digital image by means of an edge detection process, determining respective sides of each of said hard border fragments, and forming segments by determining for each pixel a closest side and by uniquely assigning each pixel of said digital image to a single segment, corresponding with the determined closest side to the respective pixel.

Abstract: The invention relates to a system (120) that makes recommendations on what to wear. Clothes (101-105) are tagged (106-110) so that taking them out of storage (100) can be detected. Using this information, a user profile is updated. This way, the system (120) learns from the user's behavior and is able to generate more accurate recommendations over time. The user profile can be stored on a smart card (320), so a user can use it in a shop to buy new clothes that fit with his current set (101-105). The user profile can also be used for targeted advertising.

Abstract: In block-based motion or depth estimation, a block is assigned a motion or depth value as a result of minimizing the matching error over a limited set (45) of candidate values (20, 21, 22, 23). When the chosen candidate value (23) is an extreme value of the set, it is impossible to predict how remote this value is from the real minimal matching error. To overcome this problem, the method and system for choosing an optimal candidate value choose a new candidate (30) beyond the extreme value (23). When this new candidate (30) then has the lowest matching error, additional new candidates are chosen until it no longer has. An apparatus for adapting a video signal (40) uses the chosen candidate values to create an enhanced version of the video signal (40).

Abstract: In block-based motion or depth estimation, a block is assigned a motion or depth value as a result of minimizing the matching error over a limited set of candidate values (21, 22, 23, 24). The curvature (20) of the function of matching error against candidate value, based on this set, can be used as a measure for the strength of the obtained optimal candidate value for motion or depth. If the obtained value is not sufficiently strong, the method and system according to the invention will repeat the process of determining the optimal candidate value until a sufficiently strong value is obtained. An apparatus for adapting a video signal (40) uses the chosen candidate values to create an enhanced version of the video signal (40).