Research on SSD for the small target detection method based on deep supervision

Taotao Zhao

doi:10.1117/12.3012264

21 December 2023 Research on SSD for the small target detection method based on deep supervision

Taotao Zhao

Proceedings Volume 12970, Fourth International Conference on Signal Processing and Computer Science (SPCS 2023); 1297037 (2023) https://doi.org/10.1117/12.3012264
Event: Fourth International Conference on Signal Processing and Computer Science (SPCS 2023), 2023, Guilin, China

Abstract

SSD is a single-stage target detection model that uses different scales feature maps to make predictions. However, due to the insufficient nonlinearity of shallow feature maps and the lack of semantic information, the detection effect of small targets is much lower than that of medium and large targets. To address this issue, this paper proposes a strategy of deep supervision and feature fusion method based on the idea of a feature pyramid network on the basis of the SSD. Through a deep supervision strategy, the same supervisory signal is applied to multiple scales of features before fusion, narrowing the semantic gap between different scale features and improving the representation ability of multi-scale features for better fusion. The top-down fusion of features extracted from the SSD backbone network is used to improve the problem of lack of semantic information in shallow feature maps. The experimental results on the NWPUVHR-10 remote sensing dataset show that the average detection accuracy for small target detection is 88.3%, higher than the average detection accuracy of DSSD (78.7%), FSSD (86.7%), and SSD (84.2%).

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Taotao Zhao "Research on SSD for the small target detection method based on deep supervision", Proc. SPIE 12970, Fourth International Conference on Signal Processing and Computer Science (SPCS 2023), 1297037 (21 December 2023); https://doi.org/10.1117/12.3012264

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
5 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Target detection

Feature fusion

Small targets

Semantics

Object detection

Feature extraction

Performance modeling

Keywords/Phrases

Search In:

Publication Years