Skip to content

Latest commit

 

History

History
37 lines (22 loc) · 2.39 KB

File metadata and controls

37 lines (22 loc) · 2.39 KB

Video Denoising using Low Rank Matrix Completion

Table of Contents

About

This project was done as part of the course CS 754: Advanced Image Processing at Indian Institute of Technology, Bombay. In this project, we have implemented a video denoising algorithm using Low Rank Matrix Completion. The algorithm is based on the paper Robust video denoising using low rank matrix completion.

The results obtained from this project are shown in the following figure.

Noisy video Median filtered video Denoised video
noisy-video-param-set1_4ZEpXnwz median-filtered-video-param-set-1_7pjAAWy1 denoised-video-param-set-1_Jj9Hr5fc
noisy-video-param-set1_r75ih9jh median-filtered-video-param-set-1_0AkW8BA8 denoised-video-param-set-1_ZDGQFTRi

Getting Started

All codes written as part of this project are tested on Matlab R2020b. To run these codes a Matlab compiler is required. Apart from basic Matlab setup, no extra packages are required.

For reading the videos written in '.y4m' format, 'yuv4mpeg2mov' package has been used. This package is included in the Implementations folder.

Usage

To reproduce the results, run the code in 'Implementaions/main.m'. The code will use the "Bus" dataset included in the Data folder by default. To denoise any other video, change the path to the video in the 'main.m' file. The results obtained will be stored in the 'Results' folder.

Contributing

Team members: Vinit Awale Piyush Bharambe