AnimalFormer: Multimodal Vision Framework for Behavior-based Precision Livestock Farming

TL;DR

This paper presents a multimodal vision framework that combines multiple AI models to analyze livestock behavior from videos, enabling non-invasive, detailed monitoring for improved farm management and animal welfare.

Contribution

The paper introduces an integrated multimodal vision framework using GroundingDINO, HQSAM, and ViTPose for comprehensive, non-invasive livestock behavior analysis from video data.

Findings

Accurately detects and segments individual animals in videos.

Provides detailed posture and movement analysis across species.

Applicable to various video resolutions and livestock behaviors.

Abstract

We introduce a multimodal vision framework for precision livestock farming, harnessing the power of GroundingDINO, HQSAM, and ViTPose models. This integrated suite enables comprehensive behavioral analytics from video data without invasive animal tagging. GroundingDINO generates accurate bounding boxes around livestock, while HQSAM segments individual animals within these boxes. ViTPose estimates key body points, facilitating posture and movement analysis. Demonstrated on a sheep dataset with grazing, running, sitting, standing, and walking activities, our framework extracts invaluable insights: activity and grazing patterns, interaction dynamics, and detailed postural evaluations. Applicable across species and video resolutions, this framework revolutionizes non-invasive livestock monitoring for activity detection, counting, health assessments, and posture analyses. It empowers data-driven farm management, optimizing animal welfare and productivity through AI-powered behavioral understanding.