Analysis and Comparison of Image Enhancement Techniques for Improving PSNR of Liver Images by Linear Contrast Algorithm over Median Filtering

Abstract

Aim: The goal of this research is to reduce noise present in the liver images in order to enhance it using linear contrast and median filters. And also to analyze output of both the filters based on its Peak Signal to Noise Ratio (PSNR). Materials and Methods: The research includes two groups; each group has a sample size of 20. Grayscale medical images collected from the kaggle website were used in this research. Samples were considered as (N=20) for guided filter and (N=20) for fast bilateral filter with total sample size 40 calculated using clinicalc.com. For this study, the affected and normal liver images were collected from the Kaggle website. Then the linear filtering, median filtering algorithms were executed using Matlab software. Sample size was calculated using clinicalc.com, and the comparison analysis has been carried out through SPSS software. This research contains two groups, with a Gpower of 80 percent. The performance of the novel median filter is evaluated and the performance measure PSNR is compared with the linear contrast filter. Result: Based on Matlab simulation results, the PSNR of novel median filters is 76.0355 and linear contrast filters have PSNR of 57.1785. From the statistical analysis, it is observed that the significant value of PSNR (Peak Signal to Noise Ratio) (0.409) and p>0.05. Conclusion: This study reveals that for the medical image enhancement purpose the novel median filter provides better PSNR than the linear contrast algorithm on ultrasound liver images.