Scalable Vision System for Mouse Homecage Ethology Ghadi Salem 1(B ) , Jonathan Krynitsky 1 , Brett Kirkland 1 , Eugene Lin 1 , Aaron Chan 1 , Simeon Anfinrud 1 , Sarah Anderson 1 , Marcial Garmendia-Cedillos 1 , Rhamy Belayachi 1 , Juan Alonso-Cruz 1 , Joshua Yu 1 , Anthony Iano-Fletcher 1 , George Dold 1 , Tom Talbot 1 , Alexxai V. Kravitz 1 , James B. Mitchell 1 , Guanhang Wu 2 , John U. Dennis 2 , Monson Hayes 3 , Kristin Branson 4 , and Thomas Pohida 1 1 National Institutes of Health, Bethesda, MD, USA ghadi.salem@nih.gov 2 Food and Drug Administration, Silver Spring, MD, USA 3 George Mason University, Fairfax, VA, USA 4 Howard Hughes Medical Institute (JFRC), Ashburn, VA, USA Abstract. In recent years, researchers and laboratory support compa- nies have recognized the utility of automated profiling of laboratory mouse activity and behavior in the home-cage. Video-based systems have emerged as a viable solution for non-invasive mouse monitoring. Wider use of vision systems for ethology studies requires the develop- ment of scalable hardware seamlessly integrated with vivarium venti- lated racks. Compact hardware combined with automated video analysis would greatly impact animal science and animal-based research. Auto- mated vision systems, free of bias and intensive labor, can accurately assess rodent activity (e.g., well-being) and behavior 24-7 during research studies within primary home-cages. Scalable compact hardware designs impose constraints, such as use of fisheye lenses, placing greater bur- den (e.g., distorted image) on downstream video analysis algorithms. We present novel methods for analysis of video acquired through such spe- cialized hardware. Our algorithms estimate the 3D pose of mouse from monocular images. We present a thorough examination of the algorithm training parameters’ influence on system accuracy. Overall, the methods presented offer novel approaches for accurate activity and behavior esti- mation practical for large-scale use of vision systems in animal facilities. 1 Introduction The application of vision systems’ technologies could have a huge impact on animal-based medical research, including corresponding animal care. During recent decades, the use of laboratory mice in biomedical research increased con- siderably [1]. Laboratory animals including mice are used to gain new knowledge for improving the health and well-being of both humans and other animals [2]. The rights of this work are transferred to the extent transferable according to title 17 U.S.C. 105. c  Springer International Publishing AG 2016 J. Blanc-Talon et al. (Eds.): ACIVS 2016, LNCS 10016, pp. 626–637, 2016. DOI: 10.1007/978-3-319-48680-2 55