An optimal adaptive reweighted sampling-based adaptive block compressed sensing for underwater image compression

The use of Block Compressed Sensing (BCS) as an alternative to conventional Compressed Sensing (CS) in image sampling and acquisition has gained attention due to its potential benefits. However, BCS can suffer from blocking artifacts and blurs in the reconstructed images, which can degrade the overa...

Full description

Saved in:
Bibliographic Details
Published in:The Visual computer 2024-06, Vol.40 (6), p.4071-4084
Main Authors: Monika, R., Dhanalakshmi, Samiappan
Format: Article
Language:English
Subjects:
Adaptive sampling
Algorithms
Artificial Intelligence
Blocking
Computer Graphics
Computer Science
Cross correlation
Data compression
Image acquisition
Image compression
Image enhancement
Image Processing and Computer Vision
Image quality
Image reconstruction
Original Article
Sensors
Signal to noise ratio
Underwater
Wavelet transforms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The use of Block Compressed Sensing (BCS) as an alternative to conventional Compressed Sensing (CS) in image sampling and acquisition has gained attention due to its potential benefits. However, BCS can suffer from blocking artifacts and blurs in the reconstructed images, which can degrade the overall image quality. To address these issues and improve reconstruction performance, Adaptive Block Compressed Sensing (ABCS) techniques can be used. ABCS minimizes the blurs and artifacts that occur during the reconstruction process by adaptively selecting samples from different image blocks. To further enhance the sampling efficiency and overall performance in underwater image compression, a new approach called adaptive reweighted sampling-based ABCS (ARS-ABCS) in Fast Haar Wavelet Transform (FHWT) domain is proposed in the paper. This reweighting process allows the system to allocate more samples to the areas where reconstruction quality is low or artifacts are prevalent, improving the overall image reconstruction in a targeted manner. Performance is measured in terms of Peak Signal to Noise Ratio (PSNR), Structural SIMilarity index (SSIM), Normalized Cross-Correlation (NCC) and Normalized Absolute Error (NAE). The results demonstrate that the proposed ARS-ABCS has achieved 1.5 to 5dB increase in PSNR with respect to other non-weighted adaptive schemes. It has produced space saving of 60 to 70% with utilization of only around 30% of total samples in the image. SSIM and NCC values obtained are closer to 1 with low NAE values.
ISSN:0178-2789
1432-2315
DOI:10.1007/s00371-023-03069-5