Automated Dead Chicken Detection in Poultry Farms Using Knowledge Distillation and Vision Transformers DOI Creative Commons

Ridip Khanal,

Wenqin Wu, Joonwhoan Lee

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 15(1), P. 136 - 136

Published: Dec. 27, 2024

Detecting dead chickens in broiler farms is critical for maintaining animal welfare and preventing disease outbreaks. This study presents an automated system that leverages CCTV footage to detect chickens, utilizing a two-step approach improve detection accuracy efficiency. First, stationary regions the footage—likely representing chickens—are identified. Then, deep learning classifier, enhanced through knowledge distillation, confirms whether detected object indeed chicken. EfficientNet-B0 employed as teacher model, while DeiT-Tiny functions student balancing high computational A dynamic frame selection strategy optimizes resource usage by adjusting monitoring intervals based on chickens’ age, ensuring real-time performance resource-constrained environments. method addresses key challenges such lack of explicit annotations along with common farm issues like lighting variations, occlusions, cluttered backgrounds, chicken growth, camera distortions. The experimental results demonstrate validation accuracies 99.3% model 98.7% significant reductions demands. system’s robustness scalability make it suitable large-scale deployment, minimizing need labor-intensive manual inspections. Future work will explore integrating methods incorporate temporal attention mechanisms removal processes.

Language: Английский

Smart glasses in the chicken barn: Enhancing animal welfare through mixed reality DOI
Dorian Baltzer,

Shannon Douglas,

Jan‐Henrik Haunert

et al.

Smart Agricultural Technology, Journal Year: 2025, Volume and Issue: 10, P. 100786 - 100786

Published: Jan. 18, 2025

Language: Английский

Citations

1
DiReDi: Distillation and Reverse Distillation for AIoT Applications DOI Creative Commons
Chen Sun, Q.K. Tong, Wenshuang Yang

et al.

IEEE Open Journal of the Computer Society, Journal Year: 2024, Volume and Issue: 5, P. 748 - 760

Published: Jan. 1, 2024

Language: Английский

Citations

0
Automated Dead Chicken Detection in Poultry Farms Using Knowledge Distillation and Vision Transformers DOI Creative Commons

Ridip Khanal,

Wenqin Wu, Joonwhoan Lee

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 15(1), P. 136 - 136

Published: Dec. 27, 2024

Detecting dead chickens in broiler farms is critical for maintaining animal welfare and preventing disease outbreaks. This study presents an automated system that leverages CCTV footage to detect chickens, utilizing a two-step approach improve detection accuracy efficiency. First, stationary regions the footage—likely representing chickens—are identified. Then, deep learning classifier, enhanced through knowledge distillation, confirms whether detected object indeed chicken. EfficientNet-B0 employed as teacher model, while DeiT-Tiny functions student balancing high computational A dynamic frame selection strategy optimizes resource usage by adjusting monitoring intervals based on chickens’ age, ensuring real-time performance resource-constrained environments. method addresses key challenges such lack of explicit annotations along with common farm issues like lighting variations, occlusions, cluttered backgrounds, chicken growth, camera distortions. The experimental results demonstrate validation accuracies 99.3% model 98.7% significant reductions demands. system’s robustness scalability make it suitable large-scale deployment, minimizing need labor-intensive manual inspections. Future work will explore integrating methods incorporate temporal attention mechanisms removal processes.

Language: Английский

Citations

0