Image Quality Enhancement System for Medical Management Applications

EasyChair Preprint no. 10451

Abstract

Medical imaging that is accurate and of a high standard is crucial to the field of healthcare for the proper diagnosis, formulation of a treatment plan, and monitoring of a variety of illnesses. However, due to inherent difficulties, getting clear and detailed photos might be difficult. In this , the bilateral filter (BF) is employed as a preprocessing step to solve the issue of noise in the photos. We then used segmentation methods based on convolutional neural networks (CNN) to precisely locate the tumor location and improve the picture. On the MRI image dataset, training, testing, and validation were performed. Images either identify tumors or don't, depending on the approach. Then image enhancement is used to enhance the quality, reduce noise, improve contrast and recover image details .Several performance criteria, including accuracy and SNR, were used to assess the work's findings. When compared to existing methods, the suggested methodology for image enhancement utilizing Resnet 50 displays greater performance with an SNR of 60 dB and an accuracy of 98%.

Keyphrases: Brain tumor., CNN (Resnet 50), Computed Tomography, Contrast enhancement, image processing, image segmentation, Retinex

@Booklet{EasyChair:10451,
  author = {B.Surekha Reddy and Akhila Gopakumar and Mohith Reddy and Chandana Rachakonda},
  title = {Image Quality Enhancement System for Medical Management Applications},
  howpublished = {EasyChair Preprint no. 10451},

  year = {EasyChair, 2023}}