计算机应用 ›› 2014, Vol. 34 ›› Issue (2): 533-537.

• 人工智能 • 上一篇    下一篇

基于人计算的小鼠行为识别

刘景,邓莎莎,童晶,陈正鸣   

  1. 河海大学 物联网工程学院,江苏 常州 213022
  • 收稿日期:2013-07-29 修回日期:2013-09-08 出版日期:2014-02-01 发布日期:2014-03-01
  • 通讯作者: 刘景
  • 作者简介:刘景(1973-),男,山东费县人,副教授,博士,CCF高级会员,主要研究方向:计算机图形学、计算机视觉;邓莎莎(1987-),女,湖北随州人,硕士研究生,主要研究方向:计算机视觉;童晶(1981-),男,江苏扬州人,讲师,博士,主要研究方向:计算机图形学、计算机视觉;陈正鸣(1965-),男,浙江东阳人,教授,博士,CCF高级会员,主要研究方向:特征造型与识别、CAD/CAM集成。
  • 基金资助:
    国家自然科学基金资助项目

Mouse behavior recognition based on human computation

LIU Jing,DENG Shasha,TONG Jing,CHEN Zhengming   

  1. College of IOT Engineering, Hohai University, Changzhou Jiangsu 213022, China
  • Received:2013-07-29 Revised:2013-09-08 Online:2014-02-01 Published:2014-03-01
  • Contact: LIU Jing

摘要: 已有的计算机自动分析系统很难准确识别小鼠行为,且普遍采用专家对大量视频图像进行行为标注的方法获得真实值,但专家标注存在一定的主观误判。针对上述问题,提出将人计算思想应用于小鼠行为识别中,让人辅助计算机识别小鼠行为。首先利用人在视觉感知等方面的优势,以及网络的分众、协同性,把人作为分布式个体处理单元,将小鼠行为视频按帧分配给网络上的人,同一帧图像由若干人进行行为分类,然后计算机系统对所有有效的分类结果进行统计、分析和处理,实现对视频图像序列的行为分类。实验表明,该方法在有限代价下,能够有效地提高小鼠行为的正确识别率。

关键词: 人计算, 小鼠行为识别, 微任务, 质量控制, 优化

Abstract: The mouse behaviors cannot be accurately recognized by the existing computer-based automatic analysis system, and the ground truth is generally achieved from experts’ annotation on a massive number of video images. However, to some extent, subjective misjudgments are unavoidable. To solve these problems, a human computation-based mouse behavior recognition method was proposed in this paper. Because of the superiority of human visual perception, and the decentralization and cooperation of the internet, human brains were treated as processors in a distributed system. Firstly, every mouse behavior frames were distributed to on-line individuals randomly, and each behavior frame was classified by a large number of on-line individuals. Secondly, all the effective classifications from the on-line individuals were collected, analyzed and processed by computer system, realizing the final mouse behavior classification based on these frame sequences. The experimental results show that the proposed method is effective to improve the correct recognition rate of mouse behaviors with limited cost.

Key words: human computation, mouse behavior recognition, micro-task, quality control, optimization

中图分类号: