Abstract: A three-dimensional image data is formulated and saved in a memory for indicating a three-dimensional shape of an object and reflectivity or color at every point of the object. For each of multiple pose candidates, an image space is created for representing brightness values of a set of two-dimensional images of the object which is placed in the same position and orientation as the each pose candidate. The brightness values are those which would be obtained if the object is illuminated under varying lighting conditions. For each pose candidate, an image candidate is detected within the image space using the 3D model data and a distance from the image candidate to an input image is determined. Corresponding to the image candidate whose distance is smallest, one of the pose candidates is selected. The image space is preferably created from each of a set of pose variants of each pose candidate.

Abstract: Three-dimensional geometric model generation method, apparatus and record medium for generating a three-dimensional geometric model indicating visual information of a three-dimensional body, comprising: simultaneously radiating light against a predetermined plane of the three-dimensional body from a frontal direction, a left and lower direction, and a right and lower direction of the above predetermined plane to uniformly illuminate the predetermined plane; photographing a predetermined plane of the three-dimensional body, to which the light was radiated, at least from one direction to acquire an image of the predetermined plane of the three-dimensional body having neither a shadow nor shading; and based on three-dimensional shape data indicating a three-dimensional shape of a predetermined plane of the three-dimensional body, and the image that was photographed, by acquiring a brightness value at each point of the three-dimensional shape of the three-dimensional body as the visual information, generating the

Abstract: A three-dimensional image data is formulated and saved in a memory for indicating a three-dimensional shape of an object and reflectivity or color at every point of the object. For each of multiple pose candidates, an image space is created for representing brightness values of a set of two-dimensional images of the object which is placed in the same position and orientation as the each pose candidate. The brightness values are those which would be obtained if the object is illuminated under varying lighting conditions. For each pose candidate, an image candidate is detected within the image space using the 3D model data and a distance from the image candidate to an input image is determined. Corresponding to the image candidate whose distance is smallest, one of the pose candidates is selected. The image space is preferably created from each of a set of pose variants of each pose candidate.

Abstract: An image comparison device for comparing a target object (an object to be recognized or identified) with registered objects (objects which are preliminarily registered to be compared with the target object) includes a registration section and a comparison section. The registration section measures the 3-D shape and surface reflectance of each registered object and registers and stores them as registered data. In the comparison section, a photographing section photographs a target object and thereby obtains an input image. A position/pose estimation section estimates the position/pose of the target object in the input image and thereby obtains position/pose parameters.

Abstract: A three-dimensional image data is formulated and saved in a memory for indicating a three-dimensional shape of an object and reflectivity or color at every point of the object. For each of multiple pose candidates, an image space is created for representing brightness values of a set of two-dimensional images of the object which is placed in the same position and orientation as the each pose candidate. The brightness values are those which would be obtained if the object is illuminated under varying lighting conditions. For each pose candidate, an image candidate is detected within the image space using the 3D model data and a distance from the image candidate to an input image is determined. Corresponding to the image candidate whose distance is smallest, one of the pose candidates is selected. The image space is preferably created from each of a set of pose variants of each pose candidate.

Abstract: 3 A 3D model storage means (170) stores the 3D shape data of a target object and illumination base data in advance. A comparison image generation means (110) generates, as a comparison image, a reproduced image with the target object being arranged in the position/posture of the current estimation value under the same illumination condition as that for the input image on the basis of the 3D shape data and illumination base data. An image displacement distribution detection means (120) segments the comparison image into partial images (sub-regions) and detects the image displacement distribution between the comparison image and the input image for each sub-region. A posture difference calculation means (130) calculates a position/posture difference value on the basis of the image displacement distribution and 3D shape data.

Abstract: An efficient computation of low-dimensional linear subspaces that optimally contain the set of images that are generated by varying the illumination impinging on the surface of a three-dimensional object for many different relative positions of that object and the viewing camera. The matrix elements of the spatial covariance matrix for an object are calculated for an arbitrary pre-determined distribution of illumination conditions. The maximum complexity is reduced for the model by approximating any pair of normal-vector and albedo from the set of all such pairs of albedo and normals with the centers of the clusters that are the result of the vector quantization of this set. For an object, a viewpoint-independent covariance matrix whose complexity is large, but practical, is constructed and diagonalized off-line. A viewpoint-dependent covariance matrix is computed from the viewpoint-independent diagonalization results and is diagonalized online in real time.

Abstract: A 3-dimensional measurement method captures a plurality of images of an object from different points of view while shifting a light pattern projected onto the object. A plurality of 3-dimensional in-sight appearances of the object are measured based on the plurality of images. A difference between measured three-dimensional in-sight appearances is corrected to finally synthesize the 3-dimensional appearance of the object.

Abstract: Three-dimensional geometric model generation method, apparatus and record medium for generating a three-dimensional geometric model indicating visual information of a three-dimensional body, comprising: simultaneously radiating light against a predetermined plane of the three-dimensional body from a frontal direction, a left and lower direction, and a right and lower direction of the above predetermined plane to uniformly illuminate the predetermined plane; photographing a predetermined plane of the three-dimensional body, to which the light was radiated, at least from one direction to acquire an image of the predetermined plane of the three-dimensional body having neither a shadow nor shading; and based on three-dimensional shape data indicating a three-dimensional shape of a predetermined plane of the three-dimensional body, and the image that was photographed, by acquiring a brightness value at each point of the three-dimensional shape of the three-dimensional body as the visual information, generating the

Abstract: A three-dimensional image data is formulated and saved in a memory for indicating a three-dimensional shape of an object and reflectivity or color at every point of the object. For each of multiple pose candidates, an image space is created for representing brightness values of a set of two-dimensional images of the object which is placed in the same position and orientation as the each pose candidate. The brightness values are those which would be obtained if the object is illuminated under varying lighting conditions. For each pose candidate, an image candidate is detected within the image space using the 3D model data and a distance from the image candidate to an input image is determined. Corresponding to the image candidate whose distance is smallest, one of the pose candidates is selected. The image space is preferably created from each of a set of pose variants of each pose candidate.

Abstract: An efficient computation of low-dimensional linear subspaces that optimally contain the set of images that are generated by varying the illumination impinging on the surface of a three-dimensional object for many different relative positions of that object and the viewing camera. The matrix elements of the spatial covariance matrix for an object are calculated for an arbitrary pre-determined distribution of illumination conditions. The maximum complexity is reduced for the model by approximating any pair of normal-vector and albedo from the set of all such pairs of albedo and normals with the centers of the clusters that are the result of the vector quantization of this set. For an object, a viewpoint-independent covariance matrix whose complexity is large, but practical, is constructed and diagonalized off-line. A viewpoint-dependent covariance matrix is computed from the viewpoint-independent diagonalization results and is diagonalized online in real time.

Abstract: An object collation method comprising a registration procedure for registering the registered data of a registered object in a database, and a collation procedure for collating the input image of a target object with the registered data. The registration procedure includes a step of storing the three-dimensional shape of the registered object and a texture space defined by a texture group indicating the luminance and/or color information of each position of the object surface under various illumination conditions. The collation procedure includes the steps of: generating an illumination fluctuation space defined by the image group under the various illumination conditions, at the location and position of the target object in the input image from the three-dimensional shape and the texture space; and collating the target object and the registered object based on the distance between the illumination fluctuation space and the input image.

Abstract: A control unit repeatedly photographs an image with cameras while phase shifting a sinusoidal light pattern from light projectors. A phase calculation unit receives image signals and outputs a phase signal from a set of images photographed with a sinusoidal light pattern of different phases. Absolute phase determining units receive two of the phase signals to determine the absolute phase value of the pixel positions. An absolute phase conversion unit receives the phase signal switched by a three-dimensional signal coordinate calculation control signal and the absolute phase signals to convert the phase value of the pixel positions into absolute phase. A three-dimensional coordinate unit receives the absolute phase conversion signal and the three-dimensional coordinate calculation control signal and outputs a three-dimensional coordinate signal.

Abstract: An image comparison device for comparing a target object (an object to be recognized or identified) with registered objects (objects which are preliminarily registered to be compared with the target object) includes a registration section and a comparison section. The registration section measures the 3-D shape and surface reflectance of each registered object and registers and stores them as registered data. In the comparison section, a photographing section photographs a target object and thereby obtains an input image. A position/pose estimation section estimates the position/pose of the target object in the input image and thereby obtains position/pose parameters.