Single-stage high performance network for tiny face detection

Junyuan He

doi:10.1117/12.2640180

1 August 2022 Single-stage high performance network for tiny face detection

Junyuan He

Proceedings Volume 12257, 4th International Conference on Information Science, Electrical, and Automation Engineering (ISEAE 2022); 122571P (2022) https://doi.org/10.1117/12.2640180
Event: 4th International Conference on Information Science, Electrical, and Automation Engineering (ISEAE 2022), 2022, Guangzhou, China

Abstract

How to improve the detectability for tiny faces is a major challenge in face detection. To solve this issue, we raise a single-stage high performance network for tiny face detection. In feature fusion, multi-scale contextual feature selection and spatial pyramid pooling modules are adopted to extract contextual information with a fixed proportion to mitigate the information loss in the highest-level feature map of feature pyramid. In addition, during training, an improved anchor compensation strategy is used to match enough anchors for tiny faces to promote the learning of tiny faces. A test is conducted in the standard data set FDDB. Our method reaches 98.5% TPR (true positive rate) when the amount of false positive faces are the same as 1000. Meanwhile, a test is conducted in the standard data set WiderFace, and the terms of AP reach 97.0%, 96.2% and 91.8% on the easy, middle and hard subset. The algorithm in the paper can improve the performance of the face detector effectively.

Citation Download Citation

Junyuan He "Single-stage high performance network for tiny face detection", Proc. SPIE 12257, 4th International Conference on Information Science, Electrical, and Automation Engineering (ISEAE 2022), 122571P (1 August 2022); https://doi.org/10.1117/12.2640180

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
6 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Facial recognition systems

Sensors

Feature extraction

Feature selection

Convolution

Head

Sensor performance

Show All Keywords

Keywords/Phrases

Search In:

Publication Years