Multi-target Detection And Tracking Technology In Medical Images

Long history of the development of multi-target detection and tracking technology, it hasalways been a research hotspot in the field of computer image processing and patternrecognition. This is because of its wide range of applications, including military, aerospace,robotics, remote sensing technology and medical technology. Multi-target detection andtracking technology has far-reaching significance of the development of computer intelligent.This article discusses the special application of the multi-target detection and trackingtechnology in medical technology-computer-aided sperm analysis system. By image analysisand processing technologies and target tracking technology, it analyzes the activity of spermin the semen samples, and ultimately gives the report of the activity of sperm in the sample.Manual analysis approach has many drawbacks, including subjectivity, poor reproducibilityand low accuracy. Therefore, computer-aided sperm analysis system is to replace the manualanalysis method. This article studies and discusses target detection and target tracking inmedical images, and then improve the existing tracking algorithm and propose a new trackingprogram. The specific work is as follows.（1） Describes the computer-aided sperm analysis system structure and systemprocesses, and summarizes the currently popular multi-target tracking.（2） Describes the multi-target detection processes in single frame image, includingthe judgment of image clarity, image graying, image denoising, thresholdsegmentation and extraction of the target area. And then codes this program.（3） Describes the processes for multi-target tracking in image sequences. Then makesimprovements on the existing target state estimation method and raises a newtarget state estimation method based on adaptive motion model. Improves theexisting nearest neighbor data association algorithm and gives a solution to solvethe conflict in the data association. Finally, uses the code to implement thisalgorithm, gives the algorithm experimental results to verify the effectiveness ofthe algorithm.