RUSSIAN VERSION

MSU Cartoon Restore Filter 2.0 beta

Filter for quality improvement of compressed animated video

MSU Graphics & Media Lab (Video Group)

Take a look at this article on the new site! Follow the link
https://videoprocessing.ai/video_filters/cartoon-restore.html

Method ideas: Dr. Dmitriy Vatolin,
Alexander Parshin
Implementation: Alexey Moiseytsev, Arsaev Marat

One of the most prominent coding artifacts in highly compressed animated video are the ringing artifacts. Ringing noise preferably manifest itself around edges due to coarse quantization. The filter is designed for quality improvement of compressed cartoons like video by deleting ringing artifacts.

Change Log

Version 2.0b
[+] Added GPU implementation.
[+] Filter performance increased

Version 1.3
[+] GUI improvement
[+] Added motion adaptive mode (filter speedup!)
[+] Added new algorithm Image Similarity (beta version)

Version 1.1
[+] First filter release

Comparison with version 1.3

for configuration: Inter Core2Duo T7300 2.0GHz, Nvidia 8600M GT, 2.0GB RAM, ForceWare 174.31
comparison

Settings

Simple mode

Filter's main window

This mode allows you select one of several presets, without additional manual filter tuning. "GPU Usage" checkbox will enable GPU mode if available. Simple mode settings:

  • Algorithm Quality - Select algorithm by quality. One can select algorithm according to desired speed/quality tradeoff.
  • Compression Artefacts - Set quantity of compression artifacts in processed cartoon.
  • Motion Adaptive - Turn this feature on to spedup algorithms in cartoons with slow motion. The same option is available in the advanced mode
  • Use GPU - This enables GPU mode, if available

Advanced mode

This mode allows configure filter manually and get best quality for particular video. For all methods "Use GPU" checkbox enables GPU mode.

Combo box "Method" allows select one of four implemented algorithms.

Simple Adaptive Filter

Filter's main window
Filter parameters:
  • Smoothness - Allows to control strength of filtration.
  • Quantization Threshold - Controls amount of ringing artifacts in source video.

Fuzzy Adaptive Filter

Filter's main window
The filtration strength depends on amount of details in each block. Left sliders correspond to blocks without details, and right sliders correspond to edge blocks.

There are some presets, which allow quickly selecting filter configuration (use "Preset" list).

Multipass Bilateral Filter

Filter's main window
Filter parameters:
  • Iterations
    If this parameter is greater, more texture details replace by flat regions.
  • Geometric Dispersion
    Allows to change filter size.
  • Photometric Dispersion
    Allows to control amount of ringing artifacts.
Also some general presets can be selected.

Similarity DB Filter (beta version)

We are still working on this algorithm quality improvement. This version can demonstrate suboptimal results!

Filter's main window
The Strength parameter of the filter sets strength of a filtration of video.
This filter requires the additional artifact.db file.

Job control and integration with AviSynth

Filter supports VirtualDub Job Control, and it can be used from AviSynth. Some examples are listed below.

Simple Adaptive Filter: 

LoadVirtualDubPlugin("MSU_cartoon_restore.vdf","MSUCartoonRestore", 0) 

clip=AVISource("D:\work\video\film.avi", false, "RGB24")

clip.ConvertToRGB32.MSUCartoonRestore("simple", Smoothness, Threshold, GPUUsage)
  • Smoothness - Integer between 1 and 15. Sets filter strength.
  • Threshold - Integer between 1 and 100. Estimates amount of ringing artifacts in source video.
  • GPUUsage - 0 or 1. This will enable GPU mode if available.
Fuzzy Adaptive Filter: 
LoadVirtualDubPlugin("MSU_cartoon_restore.vdf","MSUCartoonRestore", 0) 

clip=AVISource("D:\work\video\film.avi", false, "RGB24")

clip.ConvertToRGB32.MSUCartoonRestore("fuzzy", Strength10, Strength20, Strength30, Strength40, Strength50,  
                          Strength60, Strength70, Strength80, Strength90, Strength100, ShowBlockType)
  • Strength* - Integers between 0 and 32. Sets filter strength depends on block type.
  • ShowBlockType - 0 or 1 (optional). Turns on visualization of block types.
  • GPUUsage - 0 or 1. This will enable GPU mode if available.
Multipass Bilateral Filter: 
LoadVirtualDubPlugin("MSU_cartoon_restore.vdf","MSUCartoonRestore", 0) 

clip=AVISource("D:\work\video\film.avi", false, "RGB24")

clip.ConvertToRGB32.MSUCartoonRestore("bilateral", Iterations, GeometricDispersion, PhotometricDispersion)
  • Iterations - Integer between 1 and 5. Sets the number of passes.
  • GeometricDispersion - Integer between 0 and 75. Sets the filter size.
  • PhotometricDispersion - Integer between 0 and 200. Estimates amount of ringing artifacts in source video.
  • GPUUsage - 0 or 1. This will enable GPU mode if available.

Examples

Source frame
Source frame 		MSU Cartoon Restore
Frame processed by MSU Cartoon Restore
Source frame
Source frame 		MSU Cartoon Restore
Frame processed by MSU Cartoon Restore

Source frame
Source frame 		MSU Cartoon Restore: Simple Adaptive Filter
Simple Adaptive Filter
MSU Cartoon Restore: Fuzzy Adaptive Filter
Fuzzy Adaptive Filter 		MSU Cartoon Restore: Multipass Bilateral FIlter
Multipass Bilateral FIlter

Download

E-mail: 
Please read MSU filters FAQ before mailing.

Other resources

Video resources:

3D and stereo video
Projects on 3D and stereo video processing and analysis
MSU Video Quality Measurement tools
Programs with different objective and subjective video quality metrics implementation 		Codecs comparisons
Objective and subjective quality evaluation
tests for video and image codecs
Ext. link: x264 parameters efficiency comparison
Public MSU video filters
Here are available VirtualDub and AviSynth filters. For a given type of digital video filtration we typically develop a family of different algorithms and implementations. Generally there are also versions optimized for PC and hardware implementations (ASIC/FPGA/DSP). These optimized versions can be licensed to companies. Please contact us for details via video(at)graphics.cs.msu_ru. 		Filters for companies
We are working with Intel, Samsung, RealNetworks and other companies on adapting our filters other video processing algorithms for specific video streams, applications and hardware like TV-sets, graphics cards, etc. Some of such projects are non-exclusive. Also we have internal researches. Please let us know via video(at)graphics.cs.msu_ru if you are interested in acquiring a license for such filters or making a custom R&D project on video processing, compression, computer vision.
Video codecs projects
Different research and development
projects on video codecs 		Other
Other information
Last updated: 12-May-2022

Server size: 8069 files, 1215Mb (Server statistics)

Project updated by
Server Team and MSU Video Group

Project sponsored by YUVsoft Corp.

Project supported by MSU Graphics & Media Lab