Analysis and implementation of image processing algorithm for medical application

• Main Author: Hamiriza Firdhan Ardyasta Holil
• Other Authors: Mohammed Yakoob Siyal
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/77521
• Institution: Nanyang Technological University

Image processing is a process that utilizes algorithm to implement various functions to an image. In medical world, image processing is used to reduce the amount of noise present in medical images. This is a crucial operation as medical images produced are vulnerable to being distorted by noise during acquisition, conversion or transmission. Noise that presents in the image causes lower visual quality and the loss of important details. By using the image processing algorithm, the restoration of the quality image can be done thus resulting in accurate depiction of medical images. These medical images are very vital because doctors and physicians rely on those to analyse patient’s body or diagnose any illnesses. This project simulates the noise removal process using four filters Median, Gaussian, Non-local Means and Alpha-Trimmed. To measure the effectiveness of the filter, 4 performance measurements are used. Those are Peak Signal to Noise Ratio (PSNR), Mean Square Error (MSE), Signal to Noise Ratio (SNR), and Structural Similarity (SSIM). Noises that are examined and applied in this project are Gaussian, Salt and Pepper, Poisson and Speckle. The whole experiment will be done on 4 types of medical images. These images are X-rays, Computed Tomography (CT), Magnetic Resonance Imaging (MRI) and Ultrasound.These will be corrupted separately with the 4 noises and then 4 filters will be applied respectively for each corrupted image to restore the quality of image. The performance is calculated using the 4 performance parameters. By comparing the output value of the performance, a conclusion can be deducted on the effectiveness of particular filters under varied conditions.