Search
Abstract
Denoising videos in real-time is critical in many applications, including robotics and medicine, where varying light conditions, miniaturized sensors, and optics can substantially compromise image quality. This work proposes the first video denoising method based on a deep neural network that achieves state-of-the-art performance on dynamic scenes while running in real-time on VGA video resolution with no frame latency. The backbone of our method is a novel, remarkably simple, temporal network of cascaded blocks with forward block output propagation. We train our architecture with short, long, and global residual connections by minimizing the restoration loss of pairs of frames, leading to a more effective training across noise levels. It is robust to heavy noise following PoissonGaussian noise statistics. The algorithm is evaluated on RAW and RGB data. We propose a denoising algorithm that requires no future frames to denoise a current frame, reducing its latency considerably. The visual and quantitative results show that our algorithm achieves state-of-the-art performance among efficient algorithms, achieving from two-fold to two-orders-of-magnitude speed-ups on standard benchmarks for video denoising.
Citations
-
8
CrossRef
-
0
Web of Science
-
9
Scopus
Authors (4)
Cite as
Full text
full text is not available in portal
Keywords
Details
- Category:
- Articles
- Type:
- artykuły w czasopismach
- Published in:
-
IEEE TRANSACTIONS ON IMAGE PROCESSING
no. 31,
pages 3809 - 3824,
ISSN: 1057-7149 - Language:
- English
- Publication year:
- 2022
- Bibliographic description:
- Kopa Ostrowski P., Katsaros E., Węsierski D., Jezierska A.: BP-EVD: Forward Block-Output Propagation for Efficient Video Denoising// IEEE TRANSACTIONS ON IMAGE PROCESSING -Vol. 31, (2022), s.3809-3824
- DOI:
- Digital Object Identifier (open in new tab) 10.1109/tip.2022.3176210
- Verified by:
- Gdańsk University of Technology
seen 187 times
Recommended for you
Noise profiling for speech enhancement employing machine learning models
- K. Kąkol,
- G. Korvel,
- B. Kostek