[4H3-J-13] AI application: agriculture and forestry

[4H3-J-13-02] Calving sign detection with cattle physical feature extraction from video frames

〇Ryosuke Hyodo¹, Susumu Saito^1,2, Teppei Nakano^1,2, Makoto Akabane^1,2, Tetsunori Kobayashi¹, Tetsuji Ogawa¹ (1. Waseda University, 2. Intelligent Framework Lab)

Keywords:image recognition, calving prediction, precision livestock farming

Automatic calving sign detection from cameras can help livestock farmers with preventing fatal accidents of calf during calving. Upon deployment of such camera-based detection, the system needs to be 1) working with small data (because calving does not happen frequently), 2) robust to changing environments, and 3) explainable for reasons of the prediction results. However, these requirements are not realistic for end-to-end approaches (i.e., prediction with a single DNN). This study presents a two-stage calving prediction system, in which calving-relevant information obtained by DNN-based feature extractor is taken as inputs to another DNN-based calving sign detector. Experimental comparisons demonstrated that the developed system achieved a calving precision ratio of 81% and a calving recall ratio of 91%.