Tool for objective video quality measurements, development of a new objective metrics.

During recent years, many objective video quality metrics were developed. Lossy video compression introduces several specific artefacts, like blurring and blocking, that must be evaluated numerically. To achieve this goal, MSU Blurring and MSU Blocking measure are being developed.
To ease usage and evaluation of existing and new objective metrics we decided to develop a tool for objective video quality evaluation. MSU Video Quality Measurement tool (MSU VQM) helps in our final goal to create objective measure that will be more adequate to subjective quality than all others.

Results

MSU Blurring measure allows to evaluate blurring of any point of an image, therefore it is possible to evaluate blurring without reference (unimpaired) picture. Below you can see an example of its' visualization when the reference image is available (red color means that an impaired image is blurred, green color means that it is sharper).

Source image	Impaired image	MSU Blurring visualization

MSU Blocking measure allows to evaluate blocking without reference image. A complex algorithm is utilized, it takes into acount level of details in a region, edges of objects, information from previous frame. Below you can see an example of visualization of this metric.

Image	MSU Blocking

MSU VQM tool contains implementation of the following objective measures:

• Blurring measure
• Blocking measure
• PSNR
• VQM
• SSIM
• Delta
• MSAD
• MSE

Blurring and blocking measures are being developed by our Lab, descriptions of the other measures can be found on tool's page.

This tool was used in Second H.264 codecs comparison and in Subjective Comparison of Modern Video Codecs.

References

1. Tool page on the Internet
2. Page with description of the implemented metrics
3. Second MSU H.264 Codecs Comparison
4. MSU Subjective Comparison of Modern Video Codecs 

Team

• Dmitriy S. Vatolin, Lomonosov Moscow State University