The Evaluation Of Image Tracking Performance Based On Target Identification Rules

With the rapid development of the computer vision and the continuous progress of the digital image processing technology, the image tracking has been widely used in weaponry, security, human-computer interaction, traffic control and other aspects. With a strong practical value, the performance of the image tracing is assessed objectively and accurately.Therefore, the criteria of performance evaluation need to be studied deeply.In the paper, the target identification probability criteria is analyzed and summarized and several image tracking algorithms are studied. At the module of evaluation, different tracking algorithm is used for tracing test respectively to the image sequence. Finally, the indicator parameter is used to draw the performance evaluation curve, such as identification probability, false alarm rate, the correct rate, signal-to-noise ratio, contrast, etc. In this paper, the image tracking performance evaluation curve is evaluated based on target identification criteria and achieve to assess the performance of image tracking objectively and accurately, which laid the foundation for formulate a comprehensive, standard, intuitive image tracking performance evaluation system.