상세정보
학술대회
Towards the Perceptual Quality Enhancement of Low Bit-rate Compressed Images
- 발행일
- 202006
- 출처
- Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) 2020, pp.565-569
- 협약과제
- 20HH2300, [통합과제] 초실감 테라미디어를 위한 AV부호화 및 LF미디어 원천기술 개발, 최진수
- 초록
- In this paper, a low bit-rate compressed image quality enhancement framework is presented. A recent image/video coding method and a deep learning based quality enhancement method are integrated to improve the perceptual quality of compressed images. The proposed architecture is designed to reduce the coding artifact and restore the blurred texture details. The experimental results presents that the proposed framework yields a 33% improvement in the Perceptual Index score which is consistent with visual evaluation on a sample of results.
- KSP 제안 키워드
- Coding method, Enhancement framework, Enhancement method, Image quality enhancement, Perceptual Quality, Video coding, compressed images, deep learning(DL), low bit rate, visual evaluation