Abstract
The image segmentation refers to the extraction of region of interest and it plays a vital role in medical image processing. This work proposes multilevel thresholding based on optimization technique for the extraction of region of interest and compression of DICOM images by an improved prediction lossless algorithm for telemedicine applications. The role of compression algorithm is inevitable in data storage and transfer. Compared to the conventional thresholding, multilevel thresholding technique plays an efficient role in image analysis. In this paper, the Particle Swarm Optimization (PSO), Darwinian Particle Swarm Optimization (DPSO), and Fractional Order Darwinian Particle Swarm Optimization (FODPSO) are employed in the estimation of the threshold value. The simulation results reveal that the FODPSO-based multilevel level thresholding generate superior results. The fractional coefficient in FODPSO algorithm makes it effective optimization with fast convergence rate. The classification and blending prediction-based lossless compression algorithm generates efficient results when compared with the JPEG lossy and JPEG lossless approaches. The algorithms are tested for various threshold values and higher value of PSNR indicates the proficiency of the proposed segmentation approach. The performance of the compression algorithms was validated by metrics and was found to be appropriate for data transfer in telemedicine. The algorithms are developed in Matlab2010a and tested on DICOM CT images.
Highlights
Image segmentation refers to the process of extraction of the desired region of interest
Ghamisi et al [12] used Fractional-Order Darwinian Particle Swarm Optimization (FODPSO) for the multispectral and hyperspectral image segmentation; better results are produced when compared with Otsu multilevel thresholding technique and is fast when compared with other classical bio-inspired methods
This work presents a comparative analysis of PSO Optimization techniques for multi thresholding segmentation of abdomen CT medical images
Summary
Image segmentation refers to the process of extraction of the desired region of interest. J. Anitha et al made a comparative analysis of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) with Self Organizing Map neural network for the classification of abnormal retinal images; PSO based optimization generates better results than GA [2]. Ghamisi et al [12] used Fractional-Order Darwinian Particle Swarm Optimization (FODPSO) for the multispectral and hyperspectral image segmentation; better results are produced when compared with Otsu multilevel thresholding technique and is fast when compared with other classical bio-inspired methods. Ali et al [13] coupled Fractional-Order Darwinian Particle Swarm Optimization and Mean Shift Clustering algorithm for MRI Brain image segmentation and evaluated by metrics like Jaccard coefficient and accuracy, better results were produced when compared with Fuzzy C Means. The algorithms output, performance analysis, and conclusions are drawn
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have