# Fast Image Processing with Fully-Convolutional Networks

Source: [Fast Image Processing with Fully-Convolutional Networks](<https://www.sota2.com/research/paper/68fa639c1e209131ce9ff9d0>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T20:44:17.219Z

Authors: Qifeng Chen, Jia Xu, Vladlen Koltun

Published: 2017-09-02T22:38:13.000Z

## Abstract

We present an approach to accelerating a wide variety of image processing operators. Our approach uses a fully-convolutional network that is trained on input-output pairs that demonstrate the operator's action. After training, the original operator need not be run at all. The trained network operates at full resolution and runs in constant time. We investigate the effect of network architecture on approximation accuracy, runtime, and memory footprint, and identify a specific architecture that balances these considerations. We evaluate the presented approach on ten advanced image processing operators, including multiple variational models, multiscale tone and detail manipulation, photographic style transfer, nonlocal dehazing, and nonphotorealistic stylization. All operators are approximated by the same model. Experiments demonstrate that the presented approach is significantly more accurate than prior approximation schemes. It increases approximation accuracy as measured by PSNR across the evaluated operators by 8.5 dB on the MIT-Adobe dataset (from 27.5 to 36 dB) and reduces DSSIM by a multiplicative factor of 3 compared to the most accurate prior approximation scheme, while being the fastest. We show that our models generalize across datasets and across resolutions, and investigate a number of extensions of the presented approach. The results are shown in the supplementary video at https://youtu.be/eQyfHgLx8Dc

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Raw-to-raw translation on Nikon Z6 (test)](<https://www.sota2.com/research/sota/raw-to-raw-translation-on-nikon-z6-test.md>): Raw-to-raw translation; Nikon Z6 (test); PSNR: 18.78.

- [Raw-to-raw translation on Canon 70D (test)](<https://www.sota2.com/research/sota/raw-to-raw-translation-on-canon-70d-test.md>): Raw-to-raw translation; Canon 70D (test); PSNR: 18.96.

- [Image Enhancement on MIT-Adobe-5K DPE](<https://www.sota2.com/research/sota/image-enhancement-on-mit-adobe-5k-dpe.md>): Image Enhancement; MIT-Adobe-5K DPE; PSNR: 20.66.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1709.00643>)

[PDF](<https://arxiv.org/pdf/1709.00643.pdf>)