Related Experiment Video
Updated: May 3, 2026
10:07
Thermal Imaging to Study Stress Non-invasively in Unrestrained Birds
Published on: November 6, 2015
13.4K
PoulTrans: a transformer-based model for accurate poultry condition assessment
Jun Li1,2,3, Bing Yang1,2, Junyang Chen1
1College of Information Engineering, Sichuan Agricultural University, 46 Xinkang Road, Yucheng District, Ya'an, 625000, Sichuan Province, People's Republic of China.
Scientific Reports
|April 24, 2025
Summary
PoulTrans, a new image captioning framework, uses deep learning to automatically generate poultry status reports. This advanced system enhances poultry condition assessment for farmers.
Area of Science:
- Artificial Intelligence
- Computer Vision
- Animal Science
Background:
- Deep learning significantly improves poultry image recognition for assessing animal conditions.
- Developing intuitive decision support tools for poultry management remains a challenge.
Purpose of the Study:
- To introduce PoulTrans, an innovative image captioning framework for generating detailed poultry status reports.
- To enhance automated poultry condition assessment and decision support for farmers.
Main Methods:
- Utilized a Convolutional Neural Network (CNN) integrated with a Channel Spatial Attention Segmentation Encoder (CSA_Encoder) and a Channel Spatial Memory-Guided Transformer (CSMT).
- Incorporated visual features via CNNs into CSA_Encoder for segmented attention outputs.
- Introduced a novel PS-Loss function to optimize multi-level attention and semantic precision.
Main Results:
- PoulTrans achieved high performance scores on the PSC-Captions dataset (e.g., BLEU-4: 0.501, CIDEr: 4.927).
- The framework effectively generates detailed poultry status descriptions from images.
- Experimental analyses validated the model's effectiveness and reliability.
Conclusions:
- PoulTrans offers an advanced tool for automated poultry status generation.
- The framework enhances the digital experience and decision-making capabilities for poultry farmers.
- The developed model demonstrates significant potential for practical applications in poultry farming.
