Detection and Breed Classification of Cattle Using YOLO v4 Algorithm
Alperen Yılmaz, Gizem Nur Uzun, Mustafa Zahid GÜRBÜZ, Oğuzhan Kıvrak
Abstract
Cattle detection is related to the determination of the position of cattle in the given image. Breed classification is classification of this detected animal by using the computer vision technique. This study is carried out to evaluate the performance of cattle detection and breed classification according to specific regions on the body of cattle that differ according to their breed. You Only Look Once (YOLO) the algorithm is used for cattle detection and breed classification on the data set. To demonstrate the effectiveness of the proposed technique in cattle detection and classification, a custom data set is created with images collected over Google Images. Experimental results show the success of using the YOLO algorithm in detection and breed classification on cattle images with 92.85% accuracy.