Real-time video denoising on multicores and GPUs with Kalman-based and Bilateral filters fusion

Many embedded applications rely on video processing or on video visualization. Noisy video is thus a major issue for such applications. However, video denoising requires a lot of computational effort and most of the state-of-the-art algorithms cannot be run in real-time at camera framerate. This article introduces a new real-time video denoising algorithm for embedded platforms called RTE-VD. We first compare its denoising capabilities with other online and offline algorithms. We show that RTE-VD can achieve real-time performance (25 frames per second) for qHD video (960×540 pixels) on embedded CPUs and the output image quality is comparable to state-of-theart algorithms. In order to reach real-time denoising, we applied several high-level transforms and optimizations (SIMDization, multi-core parallelization, operator fusion and pipelining). We study the relation between computation time and power consumption on several embedded CPUs and show that it is possible to determine different frequency and core configurations in order to minimize either the computation time or the energy.

show abstract

“…We compare RTE-VD to VBM3D [12], VBM4D [13] and STMFK [15]. VBM3D is one of the most popular video denoising method.…”

Section: Denoising Efficiency Comparisonmentioning

confidence: 99%

“…There are algorithms designed for real-time denoising [15] on embedded architectures [16], but only for light noise like compression artifacts, which give poor results on heavy noise.…”

Section: Introductionmentioning

confidence: 99%

A New Real-Time Embedded Video Denoising Algorithm

Petreto

Romera

Lemaître

et al. 2019

2019 Conference on Design and Architectures for Signal and Image Processing (DASIP)

View full text Add to dashboard Cite

show abstract

“…Fast algorithms rely on much simpler principles which can be implemented efficiently on GPUs or FPGAs. For example [50] relies on a bilateral filter and a Kalman filter to produce a real-time video denoising algorithm. A recursive version of NL-means is proposed in [1].…”

Section: Introductionmentioning

confidence: 99%

GPU acceleration of NL-means, BM3D and VBM3D

Davy

Ehret

2020

J Real-Time Image Proc

View full text Add to dashboard Cite

Denoising is an essential part of any image or video processing pipeline. Unfortunately, due to time processing constraints, many pipelines do not consider the use of modern denoisers. These algorithms have only CPU implementations or suboptimal GPU implementations. We propose a new efficient GPU implementation of NL-means and BM3D, and, to our knowledge, the first GPU implementation of the video denoising algorithm VBM3D. The performance of these implementations enable their use in real-time scenarios.

show abstract

“…Noise reduction is an important first step of all image pipelines and therefore chosen as an example in this work. GPU-based video denoising has also been demonstrated with speed gains [6]. However, the application to neutron imaging has not been shown.…”

Section: Introductionmentioning

confidence: 99%

GPU Accelerated Image Processing in CCD-Based Neutron Imaging

2018

View full text Add to dashboard Cite

Image processing is an important step in every imaging path in the scientific community. Especially in neutron imaging, image processing is very important to correct for image artefacts that arise from low light and high noise statistics. Due to the low global availability of neutron sources suitable for imaging, the development of these algorithms is not in the main scope of research work and once established, algorithms are not revisited for a long time and therefore not optimized for high throughput. This work shows the possible speed gain that arises from the usage of heterogeneous computing platforms in image processing along the example of an established adaptive noise reduction algorithm.

show abstract

Real-time video denoising on multicores and GPUs with Kalman-based and Bilateral filters fusion

Cited by 9 publications

References 29 publications

A New Real-Time Embedded Video Denoising Algorithm

A New Real-Time Embedded Video Denoising Algorithm

GPU acceleration of NL-means, BM3D and VBM3D

GPU Accelerated Image Processing in CCD-Based Neutron Imaging

Contact Info

Product

Resources

About