Research on object detection methods based on deformable convolutional networks

Jianwei Guo

doi:10.1117/12.3055890

9 January 2025 Research on object detection methods based on deformable convolutional networks

Jianwei Guo

Author Affiliations +

Proceedings Volume 13486, Fourth International Conference on Computer Vision, Application, and Algorithm (CVAA 2024); 134861T (2025) https://doi.org/10.1117/12.3055890
Event: Fourth International Conference on Computer Vision, Application, and Algorithm (CVAA 2024), 2024, Chengdu, China

Abstract

Object detection is a core problem in the field of computer vision and finds extensive applications in areas such as autonomous driving and security surveillance. Traditional YOLO-based object detection algorithms often struggle with complex backgrounds, varying object shapes, and detecting small objects. To address these issues, this paper proposes an improved YOLO model based on deformable convolutions. Deformable convolutional networks enhance the model’s ability to perceive complex object shapes by introducing spatial deformation capabilities. Experiments were conducted using the publicly available COCO128 dataset and the experimental results show that after introducing deformable convolutions in deeper layers of the model, overall detection accuracy improves, with the mAP50-95 reaching 62.5% when replacing the seventh layer, an increase of 1.2 percentage points compared to the original model. The results indicate that the YOLO model based on deformable convolutions offers certain advantages in handling complex object detection tasks.

(2025) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Jianwei Guo "Research on object detection methods based on deformable convolutional networks", Proc. SPIE 13486, Fourth International Conference on Computer Vision, Application, and Algorithm (CVAA 2024), 134861T (9 January 2025); https://doi.org/10.1117/12.3055890

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
8 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Object detection

Convolution

Deformation

Computer vision technology

Machine learning

Performance modeling

Feature extraction

Show All Keywords

Keywords/Phrases

Search In:

Publication Years