Coarse representation of visual object's shape for search/query/filtering applications
View Patent ↗A method of coarse representation of a visual object's shape for search/query/filtering applications uses a binding box that fully encompasses the object of interest within the image to extract a feature vector. Once the feature vector is available, matching based on specific queries may be performed using a search engine to compare the query number to an appropriate element of the feature vector, performing sorting to pick the best matches.
1. A method of coarse representation of a shape of a visible object in a digital picture comprising the steps of:
segmenting visible objects from the digital picture:
extracting a bitmap for an object of interest from the segmented visible objects; and
estimating from the bitmap a display aspect ratio and a binding box for the object of interest, wherein said estimation step additional comprises
estimating in pixel units a set of parameters for the binding box; and
normalizing the pixel units to form a feature vector representing the binding box;
searching a video database having visible objects, each visible object having an associated feature vector, as to find those objects whose feature vectors match the feature vector of the object of interest; wherein said searching step additionally comprises
computing aspect ratios for all visible objects in the video database;
computing distances according to a specified distance metric between the desired aspect ratio and the aspect ratios for the visible objects in the video database;
sorting the distances in descending order to produce a sort list of aspect ratios and associated visible objects; and
displaying the visible objects associated with the aspect ratios that are at the top of the sort list.