1. Science
  2. Publications
  3. Information Processing Systems
  4. 2(100)'2012
  5. Clustering and neural networks in no-reference image visual quality metrics

Clustering and neural networks in no-reference image visual quality metrics

O.I. Ieremeiev
Annotations languages:

The task of sharing of a few measures of visual quality of images is considered without a standard. The proper methods are developed on the basis of clusterization and artificial neuron networks. By the base of test images of NRTID верификация of separate and synthesized measures of visual quality is conducted. The best accordance perception of man of sharing of measures of visual quality is confirmed. The value of grade correlation of Spirmena for a clusterization and neuron network grew to 0,79 and 0,75 accordingly.
Keywords: measures of visual quality without a standard, system of visual perception of man, analysis of images, neuron networks, clusterization