Automated estimation of pose features in broilers using computer vision

Autor:	Fodor, I., Doornweerd, J.E., de Klerk, B., Bouwman, A.C.
Jazyk:	angličtina
Rok vydání:	2022
Předmět:	WIAS Life Science Fokkerij en Genomica Animal Breeding and Genomics Fokkerij & Genomica Animal Breeding & Genomics
Zdroj:	Proceedings of 12th World Congress on Genetics Applied to Livestock Production (WCGALP) Proceedings of 12th World Congress on Genetics Applied to Livestock Production (WCGALP). Wageningen: Wageningen Academic Publishers
Popis:	The importance of sensor-based phenotyping is increasing in the broiler industry, aiming at improved animal health and welfare, and lower economic losses. We analysed 11 pose features of 87 individual chickens at three ages (day 14, 21, and 33) using video recordings filmed from behind while the broilers walked through a corridor one by one. A pre-trained deep learning model was trained on a limited number of frames (n=181) to adapt it to a new environment for accurate keypoint detection. Extraction of the three poses of interest (double support, left and right steps) was fully automated. Significant (p
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=dedup_wf_001::20972c3bd030fe84c3899bdbdfc4ccb2 https://research.wur.nl/en/publications/automated-estimation-of-pose-features-in-broilers-using-computer- Zobrazit plný text záznamu