Object recognition with co-occurrence histograms and false alarm probability analysis for choosing optimal object recognition process parameters

摘要

This invention is directed toward an object recognition system and process that identifies the location of a modeled object in a search image. This involves first capturing model images of the object whose location is to be identified in the search image. A co-occurrence histogram (CH) is then computed for each model images. A model image CH is computed by generating counts of every pair of pixels whose pixels exhibit colors that fall within the same combination of a series of pixel color ranges and which are separated by a distance falling within the same one of a series of distance ranges. Next, a series of search windows, of a prescribed size, are generated from overlapping portions of the search image. A CH is also computed for each of these search windows using the pixel color and distance ranges established for the model image CHs. A comparison between each model image CH and each search window CH is conducted to assess their similarity. A search window that is associated with a search window CH having a degree of similarity to one of the model image CHs which exceeds a prescribed search threshold is designated as potentially containing the object being sought. This designation can be presumed final, or further refined. This system and process requires that the size of the search window, color ranges and distance ranges be chosen ahead of time. The choice of these parameters can be optimized via a false alarm analysis.