Face Image Restoration Method Using Semantic and Transformer Splitting Networks
- Authors
- Choi, Hyoungki; Choi, Jinsol; Lim, Heunseung; Paik, Joonki
- Issue Date
- Jan-2024
- Publisher
- Institute of Electrical and Electronics Engineers Inc.
- Keywords
- face restoration; generative adversarial networks; transformer
- Citation
- Digest of Technical Papers - IEEE International Conference on Consumer Electronics, v.2024 IEEE
- Journal Title
- Digest of Technical Papers - IEEE International Conference on Consumer Electronics
- Volume
- 2024 IEEE
- URI
- https://scholarworks.bwise.kr/cau/handle/2019.sw.cau/73059
- DOI
- 10.1109/ICCE59016.2024.10444243
- ISSN
- 0747-668X
- Abstract
- This paper delves into the hardware constraints of consumer-grade surveillance camera systems, proposing a unique network architecture that splits into four distinct branches tailored for mainstream consumer electronics. While there have been significant advancements in consumer camera technology, the financial barriers related to surveillance applications in consumer markets remain notably high. Responding to this, our research presents a state-of-the-art method, optimized for everyday consumer devices, to enhance facial regions in videos by utilizing our specialized splitting network design. This model, ideal for consumer technology applications, demonstrates the capacity to precisely reconstruct damaged facial features at a pixel-level, all the while preserving the true aesthetics and authenticity of human faces. Recognizing the critical role of facial regions for personal safety in consumer settings, our solution presents a compelling answer to current challenges. This research accentuates the profound potential of advanced deep learning techniques to fortify personal safety in the modern consumer electronics landscape. © 2024 IEEE.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - Graduate School of Advanced Imaging Sciences, Multimedia and Film > Department of Imaging Science and Arts > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.