This repository implements DnCNN (“Deep CNN Denoiser”) from the paper “Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising”. DnCNN is a feedforward convolutional neural network that learns to predict the residual noise (i.e. noise map) from a noisy input image, which is then subtracted to yield a clean image. This formulation allows efficient denoising, supports blind Gaussian noise (i.e. unknown noise levels), and can be extended to related tasks like image super-resolution or JPEG deblocking in some variants. The repository includes training code (using MatConvNet / MATLAB), demo scripts, pretrained models, and evaluation routines. Single model handling multiple noise levels.

Features

  • Residual learning (predicting noise rather than clean image)
  • Batch normalization to stabilize training
  • Single model handling multiple noise levels (blind denoising)
  • Demo / test scripts included
  • Pretrained model weights for ease of use
  • Extensions to super-resolution / deblocking tasks

Project Samples

Project Activity

See All Activity >

Follow DnCNN

DnCNN Web Site

Other Useful Business Software
Compliant and Reliable File Transfers Backed by Top Security Certifications Icon
Compliant and Reliable File Transfers Backed by Top Security Certifications

Cerberus FTP Server delivers SOC 2 Type II certified security and FIPS 140-2 validated encryption.

Stop relying on non-certified, legacy file transfer tools that creak under the weight of modern security demands. Get full audit trails, advanced access controls and more supported by an award-winning team of experts. Start your free 25-day trial today.
Start Free Trial
Rate This Project
Login To Rate This Project

User Reviews

Be the first to post a review of DnCNN!

Additional Project Details

Operating Systems

Linux, Mac, Windows

Programming Language

MATLAB

Related Categories

MATLAB Computer Vision Libraries

Registered

2025-09-29