FFDNet: Toward a Fast and Flexible Solution for CNN based Image Denoising

11 Oct 2017  ·  Kai Zhang, WangMeng Zuo, Lei Zhang ·

Due to the fast inference and good performance, discriminative learning methods have been widely studied in image denoising. However, these methods mostly learn a specific model for each noise level, and require multiple models for denoising images with different noise levels. They also lack flexibility to deal with spatially variant noise, limiting their applications in practical denoising. To address these issues, we present a fast and flexible denoising convolutional neural network, namely FFDNet, with a tunable noise level map as the input. The proposed FFDNet works on downsampled sub-images, achieving a good trade-off between inference speed and denoising performance. In contrast to the existing discriminative denoisers, FFDNet enjoys several desirable properties, including (i) the ability to handle a wide range of noise levels (i.e., [0, 75]) effectively with a single network, (ii) the ability to remove spatially variant noise by specifying a non-uniform noise level map, and (iii) faster speed than benchmark BM3D even on CPU without sacrificing denoising performance. Extensive experiments on synthetic and real noisy images are conducted to evaluate FFDNet in comparison with state-of-the-art denoisers. The results show that FFDNet is effective and efficient, making it highly attractive for practical denoising applications.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Grayscale Image Denoising BSD68 sigma15 FFDNet PSNR 31.63 # 11
Grayscale Image Denoising BSD68 sigma25 FFDNet PSNR 29.19 # 11
Grayscale Image Denoising BSD68 sigma35 FFDNet PSNR 27.73 # 1
Grayscale Image Denoising BSD68 sigma50 FFDNet PSNR 26.29 # 12
Grayscale Image Denoising BSD68 sigma75 FFDNet PSNR 24.79 # 1
Color Image Denoising CBSD68 sigma15 FFDNet PSNR 33.87 # 6
Color Image Denoising CBSD68 sigma25 FFDNet PSNR 31.21 # 5
Color Image Denoising CBSD68 sigma35 FFDNet PSNR 29.58 # 3
Color Image Denoising CBSD68 sigma50 FFDNet PSNR 27.96 # 7
Color Image Denoising CBSD68 sigma75 FFDNet PSNR 26.24 # 3
Grayscale Image Denoising Clip300 sigma15 FFDNet-Clip PSNR 31.68 # 1
Grayscale Image Denoising Clip300 sigma25 FFDNet-Clip PSNR 29.25 # 1
Grayscale Image Denoising Clip300 sigma35 FFDNet-Clip PSNR 27.75 # 1
Grayscale Image Denoising Clip300 sigma50 FFDNet-Clip PSNR 26.25 # 1
Grayscale Image Denoising Clip300 sigma60 FFDNet-Clip PSNR 25.51 # 1
Color Image Denoising Kodak25 sigma15 FFDNet PSNR 34.63 # 1
Color Image Denoising Kodak25 sigma25 FFDNet PSNR 32.13 # 1
Color Image Denoising Kodak25 sigma35 FFDNet PSNR 30.57 # 1
Color Image Denoising Kodak25 sigma50 FFDNet PSNR 28.98 # 1
Color Image Denoising Kodak25 sigma75 FFDNet PSNR 27.27 # 1
Color Image Denoising McMaster sigma15 FFDNet PSNR 34.66 # 1
Color Image Denoising McMaster sigma25 FFDNet PSNR 32.35 # 1
Color Image Denoising McMaster sigma35 FFDNet PSNR 30.81 # 1
Color Image Denoising McMaster sigma50 FFDNet PSNR 29.18 # 3
Color Image Denoising McMaster sigma75 FFDNet PSNR 27.33 # 1
Grayscale Image Denoising Set12 sigma15 FFDNet PSNR 25.49 # 6
Color Image Denoising urban100 sigma15 FFDNet Average PSNR 33.83 # 4

Results from Other Papers


Task Dataset Model Metric Name Metric Value Rank Source Paper Compare
Denoising Darmstadt Noise Dataset FFDNet PSNR 34.40 # 6

Methods