METHOD FOR OBJECT RECOGNITION AND APPARATUS THEREOF

摘要

Disclosed are an object recognition method and an object recognition device using the same that allow a feature used for object recognition to be selected by the use of a weighted value of the feature for each image pixel position. The object recognition device according to the present invention includes: a feature extraction unit extracting a first feature value for each pixel with respect to a positive learning image as an object recognition target and a second feature value for each pixel with respect to a negative learning image as a control group associated with the positive input image; a feature value superposition unit superposing the first feature value and the second feature value extracted for each pixel for each corresponding position; a feature differential calculation unit calculating a difference between the first feature value and the second feature value superposed for each corresponding position; and a feature screening unit screening a position of the feature to be used in the object recognition based on a calculated difference magnitude.