TP-MVCC: Tri-plane Multi-view Fusion Model for Silkie Chicken Counting

2025-09-29

AI Summary

In densely populated poultry farming scenarios, accurate counting of silky-feathered black-boned chickens is severely hindered by heavy occlusion and the inherent limitations of single-view imaging. To address this, we propose a tri-planar multi-view density estimation framework: via geometric camera calibration and projective mapping, features from multiple viewpoints are jointly aligned onto three semantically meaningful horizontal planes—ground, mid-level, and top. A spatial feature alignment module ensures cross-view consistency, while fused tri-planar features generate a high-fidelity, scene-level density map. We further introduce MV-SilkyChicken, the first real-world, multi-view dataset of silky chickens captured in operational farming environments. Extensive experiments demonstrate that our method achieves 95.1% counting accuracy in practical deployments—substantially outperforming both single-view baselines and conventional feature fusion approaches. This work delivers a robust, end-to-end, and deployable solution for intelligent livestock management.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Accurate animal counting is essential for smart farming but remains difficult in crowded scenes due to occlusions and limited camera views. To address this, we propose a tri-plane-based multi-view chicken counting model (TP-MVCC), which leverages geometric projection and tri-plane fusion to integrate features from multiple cameras onto a unified ground plane. The framework extracts single-view features, aligns them via spatial transformation, and decodes a scene-level density map for precise chicken counting. In addition, we construct the first multi-view dataset of silkie chickens under real farming conditions. Experiments show that TP-MVCC significantly outperforms single-view and conventional fusion comparisons, achieving 95.1% accuracy and strong robustness in dense, occluded scenarios, demonstrating its practical potential for intelligent agriculture.

Problem

• Counting chickens in crowded farm scenes with occlusions
• Integrating multiple camera views for accurate animal counting
• Overcoming limited perspectives in dense agricultural environments

Innovation

• Tri-plane fusion integrates multi-view camera features
• Geometric projection aligns features to ground plane
• Decodes scene-level density map for counting