Please use this identifier to cite or link to this item:
|Title:||Investigation of noise removal algorithms for medical application||Authors:||Bansal Anmol||Keywords:||DRNTU::Engineering::Electrical and electronic engineering||Issue Date:||2015||Abstract:||Noise is any undesired signal which contaminates an image. All recording devices, whether analog or digital, encompass certain characteristics which make them vulnerable to noise. An important aspect in digital image processing is the removal of noises from images. Noise reduction is the process of removing noise from a signal. A fundamental aspect of an effective image denoising model is that it can remove noise while preserving edges. In medical imaging, the need for removal of noise is very vital as noise in X-Ray images or in other medical images such as CT scan, MRI etc. may lead to improper diagnosis of the problem. The challenging problem is to effectively process and analyse such images. The results of such analysis lead to subsequent extraction, measurement and interpretation of the information. This helps the medical practitioners to gain a better insight into the structural and functional elements of the organs being imaged. Sophisticated image processing methods are required. To date, several noise removal algorithms for images have been developed, ranging from traditional image noise removal algorithms using clustering methods to applying noise filter algorithms using advanced techniques. This report will be focusing on removing noise from images using filters. There is a need to compare the performance of these noise removing filters in order to identify the most appropriate, optimal and effective filter for each type of noise for a specific kind of medical image. This report reviews the performance of 4 different types of filters (Median, Ideal, Butterworth and Wavelet). These filters are applied on medical images containing 4 types on noises (Speckle, Salt & Pepper, Gaussian and Poisson). The medical images used are of FMRI, X-Ray and Metastatic Cancer images. The experiment covered a total of 80 distinct cases. The performance of these filters will be measured and evaluated on the basis of performance metrics Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR) in order to compare the filters and thus select and identify the most optimal and efficient filter for a particular noise and image.||URI:||http://hdl.handle.net/10356/64622||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Student Reports (FYP/IA/PA/PI)|
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.