METHOD FOR SHAPE-BASED OBJECT RECOGNITION WITH MODIFIED SHAPE CONTEXT

摘要

ABSTRACT Exemplary embodiments of the present disclosure are directed towards a method for shape-based object recognition with Modified Shape Context (MSC) descriptor, the method comprises partitioning of information into different tracks and sectors by using log polar transformation which is easier to analyze local properties by an input object unit, the input object unit represented with the evaluation of the farthest distance unit, the angular measurement unit estimates the angle measure based on the slope of the representation unit and the corresponding farthest distance unit, the Modified Shape Context (MSC) descriptor estimates the frequency occurrence of each angular measurement unit with the histogram bins, the Fourier transformation unit will be applied on the histogram bins and evaluates the invariant feature vector,testing with a training database unit, whereby the training database unit trains database to reveal shapes characteristically and obtaining the result in various stages of object recognition by a recognition result unit.