Abstract: Moving objects are detected from a temporal sequence of images captured from successive viewpoints. Image flow vectors are computed from the temporal sequence of images, for example of flow between difference images derived from the sequence. An average direction of the image flow vectors associated with displacement of image areas between a first image and a second image derived from temporal sequence of images is computed. Preferably, a method of averaging is used that is designed to suppress the effect of outliers. Regions are determined in the first image. For each of the identified regions along range of positions along the computed average direction from a position of the identified region, it is determined whether or not there is a match between first image content in the identified region in the first image and second image content in a patch located within said range of positions in the second image.

Abstract: Moving objects are detected from a temporal sequence of images captured from successive viewpoints. Image flow vectors are computed from the temporal sequence of images, for example of flow between difference images derived from the sequence. An average direction of the image flow vectors associated with displacement of image areas between a first image and a second image derived from temporal sequence of images is computed. Preferably, a method of averaging is used that is designed to suppress the effect of outliers. Regions are determined in the first image. For each of the identified regions along range of positions along the computed average direction from a position of the identified region, it is determined whether or not there is a match between first image content in the identified region in the first image and second image content in a patch located within said range of positions in the second image.

Abstract: The same person is automatically recognized in different images from his or her clothing. Color pixel values of a first and second image are captures and areas are selected for a determination whether they show the same person. First histograms of pixels area are computed, representing sums of contributions from pixels with color values in histogram bins. Each histogram bin corresponds to a combination of a range of color values and a range of heights in the areas. The ranges of color values are normalized relative to a distribution of color pixel values in areas. Furthermore, second histograms of pixels in the areas are computed, the second histograms representing sums of contributions from pixels with color values in further histogram bins. The further histogram bins are at least partly unnormalized. First and second histogram intersection scores of the first and second histograms are computed. A combined detection score is computed from the first and second histogram scores.

Abstract: Method for improving the visibly of objects and recognizing objects in a set of images recorded by one or more cameras, the images of said set of images being made from mutual different geometric positions, the method comprising the steps or recording a set or subset of images by means of one camera which is moved rather freely and which makes said images during its movement, thus providing an array of subsequent images, estimating the camera movement between subsequent image recordings, also called ego-motion hereinafter, based on features of those recorded images, registering the camera images using a synthetic aperture method, recognizing said objects.

Abstract: The same person is automatically recognized in different images from his or her clothing. Color pixel values of a first and second image are captures and areas are selected for a determination whether they show the same person. First histograms of pixels area are computed, representing sums of contributions from pixels with color values in histogram bins. Each histogram bin corresponds to a combination of a range of color values and a range of heights in the areas. The ranges of color values are normalized relative to a distribution of color pixel values in areas. Furthermore, second histograms of pixels in the areas are computed, the second histograms representing sums of contributions from pixels with color values in further histogram bins. The further histogram bins are at least partly unnormalized. First and second histogram intersection scores of the first and second histograms are computed. A combined detection score is computed from the first and second histogram scores.

Abstract: Method for improving the visibly of objects and recognizing objects in a set of images recorded by one or more cameras, the images of said set of images being made from mutual different geometric positions, the method comprising the steps or recording a set or subset of images by means of one camera which is moved rather freely and which makes said images during its movement, thus providing an array of subsequent images, estimating the camera movement between subsequent image recordings, also called ego-motion hereinafter, based on features of those recorded images, registering the camera images using a synthetic aperture method, recognizing said objects.