Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2025
|
| Online Access: | https://doi.org/10.5281/zenodo.17082366 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p>Release Notes – HMA-YOLO: Enhancing Gangue Recognition in Coal Mines with Multi-Path Attention Overview</p> <p>This repository provides the official implementation of HMA-YOLO, a lightweight object detection framework designed for gangue recognition in coal mines. The framework incorporates a multi-path large kernel attention mechanism to improve feature representation, particularly under challenging industrial conditions with densely overlapping targets and small object scales.</p> <p>This release is directly associated with the manuscript: "Enhancing Gangue Recognition in Coal Mines: A Lightweight Network with Multi-Path Attention" (submitted to The Visual Computer). Users of this code are encouraged to cite the manuscript when employing this repository in their research. Contents</p> <pre><code>Source Code Core training, validation, and inference scripts (train.py, val.py, detect.py) Model modules and supporting scripts Pretrained YOLOv9-s weights (yolov9-s-converted.pt) Configuration & Documentation requirements.txt with environment dependencies README.md and Quick Start guide for usage instructions Example configuration files for reproducibility Data A small demonstration dataset is included for testing and validation purposes. Full dataset is not publicly released at this stage due to confidentiality requirements and the ongoing nature of the project. The complete dataset will be made available once the project is finalized.</code></pre> <p>Usage</p> <pre><code>Clone the repository or download this release from Zenodo. Install dependencies via requirements.txt. Follow the Quick Start guide for dataset preparation, training, and evaluation. Example scripts are provided for sub-branch removal and custom dataset integration.</code></pre> <p>Citation</p> <p>If you use this code or dataset, please cite the associated manuscript:</p> <pre><code>Zheng Wang, Enhancing Gangue Recognition in Coal Mines: A Lightweight Network with Multi-Path Attention, submitted to The Visual Computer, 2025.</code></pre> <p>Notes</p> <pre><code>This release corresponds to the initial open-source version (v1.0). Future updates will include the full dataset and extended functionalities once project confidentiality restrictions are lifted.</code></pre>