Target tracking method fusing convolutional network features and discriminant correlation filter

Abstract

The invention discloses a target tracking method fusing convolutional network features and a discriminant correlation filter. An end-to-end lightweight network system structure is established; and theconvolutional features are trained by learning rich stream information in continuous frames, thereby improving feature representation and tracking precision. A correlation filter tracking component is constructed as a special layer in a network to track a single image block; in the tracking process, a target block and multiple background blocks are tracked at the same time; by perceiving a structural relationship between a target and surrounding background blocks, a model is built for a part with a high discrimination degree for the target and a surrounding environment; a target tracking effect is measured through a relationship between a peak sidelobe ratio and a peak value of a confidence map; and under the condition of high tracking difficulty such as large-area shielding, target shapeextreme deformation, illumination drastic change, locating is performed by automatically utilizing the discriminated background part.

Description

technical field [0001] The invention relates to a target tracking method for fusing convolution network features and discriminant correlation filters. Background technique [0002] Target tracking is a basic problem in computer vision. A common process for this problem is to input a continuous video image, initialize the object of interest with a bounding box in the first frame, and estimate the target in subsequent frames. The location of the object. Visual tracking is an important technology in computer vision, which has a wide range of applications in security protection, intelligent monitoring, human-computer interaction, and automatic control systems. [0003] In recent years, many researchers have done a lot of research on visual object tracking based on discriminant correlation filter (DCF), and made great progress. With the development of methods, existing algorithms can solve the problem of motion tracking in simple motion environment well. However, previous work...

AI Technical Summary

Problems solved by technology

Since the derivation work is carried out in the Fourier frequency domain, the present invention not only retains the characteristics of high CF efficiency, but also uses the convolution feature to improve the feature representation method of the target. The limitation of the weak structure further solves the problem of convolution features in the case of large-scale occlusion. Underperforming issues, significantly improved tracking accuracy and rate

Images