학술논문

3D Body Parts Tracking of Mouse Based on RGB-D Video from Under an Open Field
Document Type
Conference
Source
2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) Engineering in Medicine & Biology Society (EMBC), 2021 43rd Annual International Conference of the IEEE. :7252-7255 Nov, 2021
Subject
Bioengineering
Three-dimensional displays
Neuroscience
Tracking
Evolution (biology)
Pose estimation
Nose
Cameras
Language
ISSN
2694-0604
Abstract
The mouse is a valuable animal model to address the neural mechanism of higher brain function and test the pharmacodynamics of new drugs. The development of novel behavioral analysis to detect subtleties of emotion is valuable for the evolution of neuroscience research and drug discovery. 3D pose estimation is expected to contribute significantly to them. Several methods for 3D pose estimation of the mouse using optical motion capture with markers and multiple cameras have been proposed, but these methods have problems such as preparing marker sets and the influence of the markers on mouse behavior. A low-cost and simple method for markerless 3D pose estimation of the mouse using a single RGB-D (Depth) camera is proposed. As a result, the proposed method improved the accuracy of limbs tracking compared to existing limbs tracking methods. In addition, this method could track other body parts (nose, base of tail) and the center of gravity.Clinical Relevance—This study could contribute to the development of neuroscience research and drug discovery by clarifying the relationship between subtle changes in mouse behavior and emotional movements.