SEGM is a model trained on images of anthropomorphic animals and a few people. It is based on yolo8s_seg and ~140 generated images of different styles and quality. Despite the small number of images, the detector works satisfactorily. The tail and body stood out separately.