Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in Piglets

In modern intensive pig farming, accurately monitoring piglet behavior is crucial for health management and improving production efficiency. However, the complexity of existing models demands high computational resources, limiting the application of piglet behavior recognition in farming environment...

Full description

Saved in:
Bibliographic Details
Main Authors: Yizhi Luo, Kai Lin, Zixuan Xiao, Yuankai Chen, Chen Yang, Deqin Xiao
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Animals
Subjects:
Online Access:https://www.mdpi.com/2076-2615/15/11/1563
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850160957589815296
author Yizhi Luo
Kai Lin
Zixuan Xiao
Yuankai Chen
Chen Yang
Deqin Xiao
author_facet Yizhi Luo
Kai Lin
Zixuan Xiao
Yuankai Chen
Chen Yang
Deqin Xiao
author_sort Yizhi Luo
collection DOAJ
description In modern intensive pig farming, accurately monitoring piglet behavior is crucial for health management and improving production efficiency. However, the complexity of existing models demands high computational resources, limiting the application of piglet behavior recognition in farming environments. In this study, the piglet multi-behavior-recognition approach is divided into three stages. In the first stage, the LAMP pruning algorithm is used to prune and optimize redundant channels, resulting in the lightweight YOLOv8-Prune. In the second stage, based on YOLOv8, the AIFI module and the Gather–Distribute mechanism are incorporated, resulting in YOLOv8-GDA. In the third stage, using YOLOv8-GDA as the teacher model and YOLOv8-Prune as the student model, knowledge distillation is employed to further enhance detection accuracy, thus obtaining the YOLOv8-Piglet model, which strikes a balance between the detection accuracy and speed. Compared to the baseline model, YOLOv8-Piglet significantly reduces model complexity while improving detection performance, with a 6.3% increase in precision, 11.2% increase in recall, and an mAP@0.5 of 91.8%. The model was deployed on the NVIDIA Jetson Orin NX edge computing platform for the evaluation. The average inference time was reduced from 353.9 ms to 163.2 ms, resulting in a 53.8% reduction in the processing time. This study achieves a balance between model compression and recognition accuracy through the collaborative optimization of pruning and knowledge extraction.
format Article
id doaj-art-386bafbcfe33423aa2d970b975b02583
institution OA Journals
issn 2076-2615
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Animals
spelling doaj-art-386bafbcfe33423aa2d970b975b025832025-08-20T02:23:00ZengMDPI AGAnimals2076-26152025-05-011511156310.3390/ani15111563Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in PigletsYizhi Luo0Kai Lin1Zixuan Xiao2Yuankai Chen3Chen Yang4Deqin Xiao5Key Laboratory of Agricultural Equipment Technology, College of Mathematics and Informatics, South China Agricultural University, Guangzhou 510642, ChinaKey Laboratory of Agricultural Equipment Technology, College of Mathematics and Informatics, South China Agricultural University, Guangzhou 510642, ChinaCollege of Engineering, South China Agricultural University, Guangzhou 510642, ChinaKey Laboratory of Agricultural Equipment Technology, College of Mathematics and Informatics, South China Agricultural University, Guangzhou 510642, ChinaInstitute of Facility Agriculture, Guangdong Academy of Agricultural Sciences, Guangzhou 510640, ChinaKey Laboratory of Agricultural Equipment Technology, College of Mathematics and Informatics, South China Agricultural University, Guangzhou 510642, ChinaIn modern intensive pig farming, accurately monitoring piglet behavior is crucial for health management and improving production efficiency. However, the complexity of existing models demands high computational resources, limiting the application of piglet behavior recognition in farming environments. In this study, the piglet multi-behavior-recognition approach is divided into three stages. In the first stage, the LAMP pruning algorithm is used to prune and optimize redundant channels, resulting in the lightweight YOLOv8-Prune. In the second stage, based on YOLOv8, the AIFI module and the Gather–Distribute mechanism are incorporated, resulting in YOLOv8-GDA. In the third stage, using YOLOv8-GDA as the teacher model and YOLOv8-Prune as the student model, knowledge distillation is employed to further enhance detection accuracy, thus obtaining the YOLOv8-Piglet model, which strikes a balance between the detection accuracy and speed. Compared to the baseline model, YOLOv8-Piglet significantly reduces model complexity while improving detection performance, with a 6.3% increase in precision, 11.2% increase in recall, and an mAP@0.5 of 91.8%. The model was deployed on the NVIDIA Jetson Orin NX edge computing platform for the evaluation. The average inference time was reduced from 353.9 ms to 163.2 ms, resulting in a 53.8% reduction in the processing time. This study achieves a balance between model compression and recognition accuracy through the collaborative optimization of pruning and knowledge extraction.https://www.mdpi.com/2076-2615/15/11/1563pigletmulti-behavior recognitionprunedistillprecision livestock farming
spellingShingle Yizhi Luo
Kai Lin
Zixuan Xiao
Yuankai Chen
Chen Yang
Deqin Xiao
Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in Piglets
Animals
piglet
multi-behavior recognition
prune
distill
precision livestock farming
title Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in Piglets
title_full Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in Piglets
title_fullStr Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in Piglets
title_full_unstemmed Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in Piglets
title_short Collaborative Optimization of Model Pruning and Knowledge Distillation for Efficient and Lightweight Multi-Behavior Recognition in Piglets
title_sort collaborative optimization of model pruning and knowledge distillation for efficient and lightweight multi behavior recognition in piglets
topic piglet
multi-behavior recognition
prune
distill
precision livestock farming
url https://www.mdpi.com/2076-2615/15/11/1563
work_keys_str_mv AT yizhiluo collaborativeoptimizationofmodelpruningandknowledgedistillationforefficientandlightweightmultibehaviorrecognitioninpiglets
AT kailin collaborativeoptimizationofmodelpruningandknowledgedistillationforefficientandlightweightmultibehaviorrecognitioninpiglets
AT zixuanxiao collaborativeoptimizationofmodelpruningandknowledgedistillationforefficientandlightweightmultibehaviorrecognitioninpiglets
AT yuankaichen collaborativeoptimizationofmodelpruningandknowledgedistillationforefficientandlightweightmultibehaviorrecognitioninpiglets
AT chenyang collaborativeoptimizationofmodelpruningandknowledgedistillationforefficientandlightweightmultibehaviorrecognitioninpiglets
AT deqinxiao collaborativeoptimizationofmodelpruningandknowledgedistillationforefficientandlightweightmultibehaviorrecognitioninpiglets