TY - GEN
T1 - Automated Real-Time Tracking System for Socially-Housed Physically Identical Mice
AU - Wang, Yanbo
AU - Savonenko, Alena V.
AU - Etienne-Cummings, Ralph
N1 - Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - Automated behavioral and cognitive testing of multiple mice in their home environment is a challenging task requiring fast identification of identical objects for online control of environment. The realization of such tracking system is very important for drug development and neuroscience research. The significant limitation of the existing tracking systems is that the computational load is too heavy to meet the needs of online analysis. This paper focuses on development of a computationally-efficient algorithm that tracks and identifies multiple mice within the temporal constraints of frame-by-frame video analysis. Our system is able to automatically track the landmarks (noses, tails and ears) of multiple mice frame by frame and find the identity by searching the metal tags around the ear positions. The algorithm has been validated using NIR videos from standard mice housing. The average execution time for each frame is 0.012 seconds, which is consistent with the necessary requirements for online tracking and identification for high-throughput automated cognitive testing.
AB - Automated behavioral and cognitive testing of multiple mice in their home environment is a challenging task requiring fast identification of identical objects for online control of environment. The realization of such tracking system is very important for drug development and neuroscience research. The significant limitation of the existing tracking systems is that the computational load is too heavy to meet the needs of online analysis. This paper focuses on development of a computationally-efficient algorithm that tracks and identifies multiple mice within the temporal constraints of frame-by-frame video analysis. Our system is able to automatically track the landmarks (noses, tails and ears) of multiple mice frame by frame and find the identity by searching the metal tags around the ear positions. The algorithm has been validated using NIR videos from standard mice housing. The average execution time for each frame is 0.012 seconds, which is consistent with the necessary requirements for online tracking and identification for high-throughput automated cognitive testing.
UR - http://www.scopus.com/inward/record.url?scp=85124222801&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85124222801&partnerID=8YFLogxK
U2 - 10.1109/BioCAS49922.2021.9645011
DO - 10.1109/BioCAS49922.2021.9645011
M3 - Conference contribution
AN - SCOPUS:85124222801
T3 - BioCAS 2021 - IEEE Biomedical Circuits and Systems Conference, Proceedings
BT - BioCAS 2021 - IEEE Biomedical Circuits and Systems Conference, Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2021 IEEE Biomedical Circuits and Systems Conference, BioCAS 2021
Y2 - 6 October 2021 through 9 October 2021
ER -