A Review of Target Tracking Algorithm Based on Deep Neural Network
Target tracking is to predict the state of the current target according to the target context information of video sequence.Deep learning is gradually widely used in the field of target tracking.This paper elaborates on the de-velopment background of target tracking algorithms and deep learning,and reviews traditional target tracking.Based on different network task functions,target tracking algorithms based deep learning is divided into:deep learning target tracking algorithm based classification,deep learning target tracking algorithm based regression,and target tracking al-gorithm based on the combination of regression and classification.This paper selects representative target tracking algo-rithm for experiments,compares the characteristics of different algorithms.Finally,this paper analyzes the problems of current target tracking algorithms based on deep learning,and prospects the future development direction.The experi-mental results show that deep Siamese tracking networks are superior in accuracy and speed,which is becoming the ma-instream tracking algorithm framework at present.
国家科技期刊平台
NSTL
万方数据
