What Image Processing Does for Medical Applications, the 5th Online Joint Class Telkom University & MMU
What Image Processing Does for Medical Applications, the 5th Online Joint Class Telkom University & MMU. Telkom University, on the 7th anniversary (14/08/2020), in collaboration with Multimedia University Malaysia, successfully held the fifth session of the Online Joint Class. Today’s topic was similar to the 3rd session’s which is “What Image Processing Does for Medical Applications”. After previously Assoc. Prof. Agus Pratondo, Ph.D., discussed about definition of image processing and its fundamental in medical application, this time, Tee Connie, Ph.D, a senior lecturer in Faculty of Information Science and Technology at Multimedia University, Malaysia, is more on the image processing steps.
According to Dr. Connie, there are five steps of image processing, including:
- Image Acquisition
- Representation & Description
- Recognition & Interpretation
Image acquisition and pre-processing include in the low-level processing, segmentation and representation & description include in the intermediate level processing, and the rest recognition & interpretation include in the high level processing.
Image acquisition is derived from capturing the image using MRI CT Scan, X Ray, and USG to forming a digital image matrix. The image is then digitized into picture element, called PIXEL. While pre-processing is an image enhancement process which includes noise suppression, contrast enhancement, and time/space filtering. The objectives of image enhancement is to process an image so that the result is more suitable than the original image for a specific application. It can be divided into two categories, namely spatial domain methods which includes point processing and mask filtering, and frequency domain methods which includes fourier transform.
Spatial domain methods refer to the aggregate of pixels composing an image. They are also the procedures that operate directly on these pixels using point processing (single pixel filtering) and mask filtering (other neighbor filtering). Basically, a mask is a small (e.g 3×3) @-D array, in which the values of the coefficients determine the nature of the process, such as image sharpening. This approach is often referred to mask processing or filtering. There are two types of filtering methods, namely smoothing filter aiming to blur image and reduce the noise, and sharpening filter aiming to highlight fine structure. The filter used to filter noise and blur can be average filter and median filter.
At the end of the class, Dr Connie also demonstrated some study cases on image processing. The participants were assigned to find out the problems of the images showed and then find its solution. The class was so interactive by then. All participants were actively involved in the discussion guided directly by Dr. Connie. All participants also found the class interesting and the topic delivered smoothly interactive. It is also hoped that next session will be as interactive as today’s so the participants can gain more meaningful insight.(IO)***