2D Animal Skeletons Keypoint Detection: Research Progress and Future Trends

Research on two-dimensional keypoint detection within the domain of computer vision has experienced significant advancements. In contrast to the high precision and applicability achieved in human applications, the field of animal keypoint detection remains in its nascent stages of development. To explore the applications and potential of skeleton keypoint detection in areas such as animal pose estimation, behavior recognition, and intelligent breeding. To this end, we focuse on analyzing animal skeletal keypoint detection models based on deep learning technology, examining their keypoint representation forms and multi-object detection strategies. Additionally, it provides a detailed analysis of the method’s practical application details and improvements. The paper not only summarizes different model algorithms, datasets, and evaluation metrics related to animal keypoint detection but also integrates various application scenarios, highlighting distinct features under different focal points. Ultimately, this review is expected to broaden the research horizons and methodologies related to animal intelligent behavior recognition, animal welfare studies, and intelligent breeding among scholars.