KCF Object tracking method and device using improved kernelized correlation filters

摘要

The present invention discloses an object tracking method and apparatus using an improved KCF. According to the present invention, it includes a processor and a memory connected to the processor, wherein the memory generates a plurality of Gaussian blur images for an input patch including an object, and extracts features and correlation filters for each Gaussian blur image A correlation response map is calculated to determine a target position in each Gaussian blur image, and a weight update score for each Gaussian blur image is calculated using the final position of the object determined in the previous frame and the target position, An object tracking device is provided that stores program instructions executable by the processor to determine an adaptive feature weight to be applied to a next frame using the weight update score.