Medium-and-large-sized quadruped animal behavior identification method based on architecture search graph convolutional network

Abstract

A medium and large quadruped animal behavior recognition method based on a framework search graph convolutional network comprises the following steps: firstly, based on animal skeleton behavior feature extraction, aiming at medium and large quadruped animal video images in a complex field environment, using a pose estimation algorithm DeepLabCut to quickly track the positions of animal body part joint points, forming a space-time skeleton graph, and carrying out space-time feature extraction on the basis of the space-time skeleton graph; and the spatial-temporal characteristics of different behaviors of the quadruped animal are captured. Then, a plurality of space-time diagram convolution operation modules based on animal skeletons are designed, a graph-based search space is formed, residual connection, a bottleneck structure and various attention mechanisms are fused, and the network is lighter while the performance of the recognition model is improved. And then, realizing the continuity of a search space based on a microarchitecture search strategy so as to automatically search a low-cost space-time diagram convolution model for behavior identification of medium and large quadruped animals, and finally realizing the purpose of distinguishing daily behaviors of the animals, thereby having a certain application prospect.

Description

technical field [0001] The invention relates to a method for behavior recognition, in particular to a method for automatically constructing a spatiotemporal graph convolution network by using a differentiable architecture search strategy and applying it to the field of behavior recognition of medium and large tetrapod skeletons. Background technique [0002] In the prior art, behavior recognition refers to the automatic identification of behavior information of a research object through static image or video sequence information, using key technologies such as moving target detection, feature information extraction, and attitude analysis. [0003] Animal behavior recognition can facilitate scientific decision-making in animal management and is widely used in animal welfare, disease prevention, and bionics research. The research on action behavior recognition refers to designing specific feature descriptors (such as color, texture, shape, edge, space, etc.), and then forming ...

AI Technical Summary

Problems solved by technology

[0006] The problem to be solved by the present invention is that most of the existing research objects of animal behavior recognition are docile domesticated animals or laboratory animals, and the adopted methods cannot solve problems such as illumination, occlusion, and viewing angle changes in different complex scenes. Well solved, requires large datasets and is not generalizable

Images