Open AccessStudy on Spontaneous Behavior Recognition of Mice Based on Frame Stream and Feature Coordinate Matching
Author(s) -
Biao Chen,
Yuan Zhuang,
Gang Xu
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1621/1/012031
Subject(s) - frame (networking) , matching (statistics) , pattern recognition (psychology) , computer science , artificial intelligence , feature (linguistics) , cluster analysis , computer vision , sorting , object (grammar) , algorithm , mathematics , linguistics , statistics , philosophy , telecommunications
The recognition of spontaneous behavior in mice is of great significance to the biological research. It not only provides an important means for pathology, pharmacy, biological neurology, but also provides great convenience to scientific researchers. In this paper, the monitoring video of mice were taken as the research object. Based on Classical frame average method to background modeling and PBAS algorithm solved the whole detection and tracking of mice, and detection and tracking of local characteristics in mice were solved by residual neutral network (ResNet). On this basis, the spontaneous behavior recognition of mice was solved by K-Means clustering algorithm. To improve the accuracy, we proposed a method of spontaneous behavior recognition in mice based on frame stream and feature coordinate matching. The effect of recognition was intuitionistic and obvious, and met the needs of subsequent experiments on matching with large-scale neuronal spike sorting in mice.