dth-yolo

https://github.com/xvzhe834/dth-yolo

Science Score: 57.0%

This score indicates how likely this project is to be science-related based on various indicators:

✓
CITATION.cff file
Found CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
✓
DOI references
Found 3 DOI reference(s) in README
○
Academic publication links
○
Academic email domains
○
Institutional organization owner
○
JOSS paper metadata
○
Scientific vocabulary similarity
Low similarity (10.7%) to scientific vocabulary

Last synced: 10 months ago · JSON representation ·

Repository

Basic Info

Host: GitHub
Owner: xvzhe834
License: other
Language: Python
Default Branch: main
Size: 1.08 MB

Statistics

Stars: 0
Watchers: 1
Forks: 0
Open Issues: 0
Releases: 0

Created over 1 year ago · Last pushed over 1 year ago

Metadata Files

Readme Contributing License Citation

DTH-YOLO: Enhanced YOLOv8n with Dynamic Task-Aligned Head for Mouseholes Detection

Introduction

In recent years, the increasing number of mouseholes in grassland areas has accelerated desertification, leading to a decline in grassland productivity and severely impacting the economic benefits of herders. Therefore, accurately and efficiently detecting and locating mouseholes has become an urgent research problem. UAV photography, with its advantages of wide coverage and flexibility, has gradually become an effective tool for mousehole detection. However, due to varying flight altitudes causing significant changes in target scale, severe occlusion of targets, and the small size of burrows, aerial detection faces high rates of missed and false detections, posing a significant challenge to burrow identification. To address these issues, we propose the DTH-YOLO model, based on YOLOv8n. First, we replace the original detection head with the Dynamic Task-Aligned Head (DTH) to enhance feature alignment across different tasks and improve detection accuracy. Secondly, we design the C2f-RVB module to optimize the feature extraction process, significantly reducing model parameters and computational costs. Finally, we introduce the Context Guided Block (CGB) for downsampling, effectively capturing contextual information in the target area, thereby improving detection performance for small objects. Experimental results show that, on a custom aerial mousehole dataset, the DTH-YOLO model achieves an 8.5\% increase in mAP@0.5, a 5.6\% improvement in precision, a 42\% reduction in parameters, and a 22\% decrease in GFLOPs compared to the baseline YOLOv8n model. Furthermore, ablation studies demonstrate the effectiveness of each proposed module.

Installation

Install PyTorch
Env: PyTorch_2.3.1; cuda_12.1; cudnn_9.6; python_3.11.11; mmcv_2.1.0

Citation

If DTH-YOLO is useful for your research, please consider citing: ``` @ARTICLE{10908834, author={Xu, Zhe and Luo, Xiaoling and Gao, Xiaojing and Pan, Xin}, journal={IEEE Access}, title={DTH-YOLO: Enhanced YOLOv8n With Dynamic Task-Aligned Head for Mousehole Detection}, year={2025}, volume={13}, number={}, pages={44912-44927}, keywords={Feature extraction;Accuracy;YOLO;Proposals;Convolution;Rodents;Grasslands;Computational modeling;Autonomous aerial vehicles;Adaptation models;YOLOv8n;object detection;deep learning;mousehole}, doi={10.1109/ACCESS.2025.3546946}}

```

Owner

Login: xvzhe834
Kind: user

Repositories: 1
Profile: https://github.com/xvzhe834

Citation (CITATION.cff)

cff-version: 1.2.0
preferred-citation:
  type: software
  message: If you use this software, please cite it as below.
  authors:
  - family-names: Jocher
    given-names: Glenn
    orcid: "https://orcid.org/0000-0001-5950-6979"
  - family-names: Chaurasia
    given-names: Ayush
    orcid: "https://orcid.org/0000-0002-7603-6750"
  - family-names: Qiu
    given-names: Jing
    orcid: "https://orcid.org/0000-0003-3783-7069"
  title: "Ultralytics YOLO"
  version: 8.0.0
  # doi: 10.5281/zenodo.3908559  # TODO
  date-released: 2023-1-10
  license: AGPL-3.0
  url: "https://github.com/ultralytics/ultralytics"

GitHub Events

Total

Watch event: 1
Delete event: 8
Issue comment event: 13
Push event: 30
Public event: 1
Pull request event: 10
Create event: 8

Last Year

Watch event: 1
Delete event: 8
Issue comment event: 13
Push event: 30
Public event: 1
Pull request event: 10
Create event: 8

Issues and Pull Requests

Last synced: 10 months ago

All Time

Total issues: 0
Total pull requests: 5
Average time to close issues: N/A
Average time to close pull requests: 5 minutes
Total issue authors: 0
Total pull request authors: 1
Average comments per issue: 0
Average comments per pull request: 2.0
Merged pull requests: 0
Bot issues: 0
Bot pull requests: 5

Past Year

Issues: 0
Pull requests: 5
Average time to close issues: N/A
Average time to close pull requests: 5 minutes
Issue authors: 0
Pull request authors: 1
Average comments per issue: 0
Average comments per pull request: 2.0
Merged pull requests: 0
Bot issues: 0
Bot pull requests: 5

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Open Source Science