Title :
Perceptually relevant ringing region detection method
Author :
Hantao Liu ; Klomp, Nick ; Heynderickx, Ingrid
Author_Institution :
Dept. of Mediamatics, Delft Univ. of Technol., Delft, Netherlands
Abstract :
A novel approach towards automatic detection of perceived ringing regions is presented. The algorithm takes into account both the physical structure and the human visual perception of the ringing artifacts. All perceived ringing regions are explicitly captured by means of a newly proposed edge detector, followed by an efficient analysis of ringing visibility around each detected edge segment. Determining visibility is based on luminance masking and texture masking as typical for the human visual system. The proposed detection method is validated by comparing its performance with the ringing regions resulting from a psychovisual experiment.
Keywords :
edge detection; image capture; image segmentation; image texture; speech intelligibility; edge detector; edge segment detection; human visual perception system; image capture; luminance masking; physical structure; psychovisual experiment; ringing artifacts; ringing region automatic detection method; ringing visibility; texture masking; Computational modeling; Detectors; Filtering; Image coding; Image edge detection; Image segmentation; Measurement;
Conference_Titel :
Signal Processing Conference, 2008 16th European
Conference_Location :
Lausanne