Paper Info
Title
Blocking-artifact detection in frequency domain for frame-rate up-conversion.
Abstract
In this paper, a blocking-artifact detection method is proposed for motion-compensated frame-rate up-conversion (MC-FRUC) algorithms. In the MC-FRUC process, most of the notable blocking artifacts occur in unreliable motion-compensated regions. To find the unreliable motion-compensated regions, an occlusion detection (OD) method for which the neighboring motion vectors are used is applied. If an occlusion occurs, then the use of median filtering reduces any inappropriate motion-compensated errors; otherwise, the normal motion compensated interpolation (MCI) is applied. In addition, to reduce the occurrence of the blocking artifacts in the MC-FRUC algorithms, a blocking-artifact-reduction method for which the blocking-artifact strength in the frequency domain is applied to the occluded regions is utilized. The experiment results show that the proposed method shows more-favorable peak signal-to-noise ratio (PSNR) values compared with those of the conventional FRUC algorithms regarding full-HD (1920 × 1080) and WQXGA (2560 × 1600) sequences.
Year
DOI
Venue
2018
10.1007/s00779-017-1019-z
Personal and Ubiquitous Computing
Keywords
Field
DocType
Frame-rate up-conversion, Motion compensated interpolation (MCI), Occlusion detection, Blocking artifacts, Integer discrete cosine transform (DCT)
Frequency domain,Computer vision,Occlusion detection,Occlusion,Median filter,Computer science,Interpolation,Artificial intelligence,Frame rate up conversion
Journal
Volume
Issue
ISSN
22
1
1617-4917
Citations 
PageRank 
References 
0
0.34
14
Authors
2
Name
Order
Citations
PageRank
Nam-Uk Kim1257.69
Yung-Lyul Lee225929.12