Region of interest-based adaptive segmentation for image compression using hybrid Jaya–Lion mathematical approach

Abstract

This paper intends to propose a new image compression technique, which is processed under various sequences of progressions. The initial process is the image segmentation and is handled by Adaptive Active Contour Model (ACM), which divides or segments the image into two regions: ROI (Region of Interest) and non-ROI. In this, the adaptiveness of this ACM is influenced with the concept of optimization. JPEG-LS algorithm is used to handle the ROI regions, and the wavelet-based lossy compression algorithm is used to handle the non-ROI region. The result of both JPEG-LS algorithm and wavelet-based compression model is combined in terms of bit-stream combination and outputs the compressed image. Subsequently, the compressed image is subjected for image decompression, which will be the reverse process of compression. This will include the bit-stream separation, which is then separately processed under both the JPEG-LS decoding and wavelet-based decomposition for attaining the ROI and non-ROI regions. Finally, the original image is attained precisely. Further, the major contribution of this work falls in the adaptiveness under optimization. The weighting factor and maximum iteration in ACM are optimally selected and for this a new hybrid optimization model that hybridizes the concept of Jaya Algorithm and Lion Algorithm is introduced.