摘要
针对精子检测和计算机辅助的精子运动特征分析所存在的问题,分析了精子密度大、运动速度快等特点和实际应用中遇到的各种困难,比较了多种阈值分割和多目标跟踪的方法,设计了基于矩量保持和卡尔曼滤波的精子检测与跟踪算法。利用矩量保持法对精子图像进行阈值分割和卡尔曼滤波对精子运动目标进行预测。实验结果表明,该算法在检测时间和精确度上都有很大提高。
Aimed at the problems of the computer-aided sperm analysis （CASA）, sperm density, movement speed, etc are analyzed, a wide range of threshold segmentation and multi-target tracking methods are compared, the moment-preserving and Kalman Filter key technologies are analyzed, and an efficient and low-cost method for automatically detecting and tracking the moving object from sperm image sequences is presented. The algorithm improves the effectiveness and accuracy of sperm detecting.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第9期2076-2079,共4页
Computer Engineering and Design
