摘要
行为绝望模型在神经心理学、神经药理学研究方面应用较为广泛,小鼠悬尾实验是最为常见的行为绝望模型实验,该实验能够对抗抑郁药、镇静药等药物的药效进行评价。利用计算机视觉处理技术,基于图像轮廓特征设计检测算法进行实验鼠行为自动分析。相比传统的人眼观察方法,基于计算机视觉的行为分析提高了实验效率,并使得实验结果更具有客观性。结果表明,该算法能够很好地区分悬尾小鼠的"主动挣扎"与"惯性摆动"状态,从而提取各运动状态起始时刻与持续时间,与人眼观察结果进行对比,"不动状态"等运动状态提取准确率达到90%以上。同时,利用图像轮廓特征,解决了传统运动提取方法中在惯性摆动幅度较大时出现的错误识别问题,从而可以更为准确地完成小鼠行为分析。
Behavioral despair model has been widely used in neuropsychology and neuropharmacology. Mice tail suspend test （TST） is the most common behavioral despair model test,which can evaluate the pharmacodynamics of antidepressants and sedatives. By means of the computer vision processing technology,combining the full body contour and other feature points,the TST behavior analysis is achieved automatically. Compared with the traditional human-eye observation method,the behavioral analysis based on computer vision improves the experimental efficiency,and also gets more objective result. Experiment shows that the algorithm can classify the “Active struggle” and “Passive swing” state of the suspend mice so as to extract the starting time and duration of each state of motion. Compared with the observation results of human eyes,the accuracy of motion state extraction such as “immobility state” is more than 90%. At the same time,the image contour feature is used to solve the problem of error recognition caused by inertial oscillation amplitude intensively in traditional motion extraction method,finishing the mice behavior analysis more accurately.
作者
李松柏
张卫华
胡光亮
LI Song-bai;ZHANG Wei-hua;HU Guang-liang(School of Computer Science,Sichuan University,Chengdu 610065,China;School of Electronic ＆ Information Engineering,Sichuan University,Chengdu 610065,China)
出处
《计算机技术与发展》
2018年第8期6-11,共6页
Computer Technology and Development
基金
国家自然科学基金(61302028)
四川省科技支撑计划基金项目(2012RZ0005)
四川省科技创新苗子工程(16-YCG061)
关键词
小鼠悬尾实验
轮廓特征
HU不变矩
行为自动分析
mice tail suspend test
contour feature
Hu invariant moment
automatic behavioral analysis
