The heart of TekMOS is a machine learning algorithm. TekMOS decodes each frame to base-band and extracts from the luminance (Y’) component a set of features related to perceived video quality characteristics. Artifacts such as blur, noise and over-compression tiling create different variations in these features from a high-quality scene.
The white paper provides the reader with the following information:
- Introduction
- What is no-reference video-quality and is it better or worse than full-reference?
- TekMOS & Machine Learning
- How does machine learning apply to video quality assessment?
- TekMOS Algorithm
- How does it work and what aspect of machine learning does it use?
- Accuracy of Evaluations
- How accurate is TekMOS?
Complete and submit the form to download the white paper "TekMOS: A general purpose, Real-Time, No-reference Video-Quality Assessment Tool."
