A Comparison of Compression Codecs for Maritime and Sonar Images in Bandwidth Constrained Applications

• Published in: Computers (Basel) Vol. 8; no. 2; p. 32
• Main Authors: Kwan, Chiman, Larkin, Jude, Budavari, Bence, Chou, Bryan, Shang, Eric, Tran, Trac D.
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.06.2019
• Subjects: Algorithms; Bandwidths; Business metrics; Channels; Codec; Compression ratio; Daala; Data compression; error recovery; Image compression; Image transmission; JPEG2000; maritime and sonar images; perceptually lossless compression; Performance measurement; Pixels; Quality standards; Software; Sonar; Video compression; X264; X265
• ISSN: 2073-431X; 2073-431X
• DOI: 10.3390/computers8020032

Since lossless compression can only achieve two to four times data compression, it may not be efficient to deploy lossless compression in bandwidth constrained applications. Instead, it would be more economical to adopt perceptually lossless compression, which can attain ten times or more compression without loss of important information. Consequently, one can transmit more images over bandwidth limited channels. In this research, we first aimed to compare and select the best compression algorithm in the literature to achieve a compression ratio of 0.1 and 40 dBs or more in terms of a performance metric known as human visual system model (HVSm) for maritime and sonar images. Our second objective was to demonstrate error concealment algorithms that can handle corrupted pixels due to transmission errors in interference-prone communication channels. Using four state-of-the-art codecs, we demonstrated that perceptually lossless compression can be achieved for realistic maritime and sonar images. At the same time, we also selected the best codec for this purpose using four performance metrics. Finally, error concealment was demonstrated to be useful in recovering lost pixels due to transmission errors.