الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 136 |  |  |
| | | from | 136 | | |
Abstract
A comparative study presents different filtering techniques applied to medical images. The performance of these techniques investigated the problem of image degradation which might occurred during the acquisition of images, camera motion, flat-bed scanner and video images. We touch upon two dimensional images of X-ray, Magnetic Resonance Imaging (MRI), Computed Tomography (CT), three dimensional pathological and scanned images with a set of predefined noise levels. We applied spatial domain and transformed domain filtering techniques. The performance of these techniques was evaluated with respect to two measures: Signal-to-Noise Ration (SNR), and shape preservation (R). It has been found that filtering using rank order filter provided the best performance in spatial domain and wavelet transform filter provided the best performance in transformed domain.
Segmentation plays an important role in identifying anatomical areas of interest for diagnosis and surgery. The segmentation process was applied to differentiate organs with various components and analyzed them.. The best techniques of the work is a sub volume, marker controlled watershed segmentation and detecting a cell using image segmentation.
Fusion of images is the process of combining two or more images into a single image retaining important features from each. We evaluated the performance of a set of image fusion techniques: averaging based, principal component analysis, discrete cosine transform (DCT), discrete wavelet transform (DWT) and artificial neural network (ANN). These techniques had been applied to a set of multi focus images to increase the clarity of the fused image. The performance of these techniques was compared using a set of indices: entropy, mean square error and peak signal to noise ratio. It was observed that the DWT fusion technique presents the most better result followed by ANN technique.
We classified unknown images and differentiate them into normal and abnormal image using Neural Network techniques, So Specialists can make a diagnostic decision by viewing specimens and measuring various diagnostically important attributes of the classified images. Two basic methods were implemented to produce the most features that applied to the architecture of ANNs based classifiers. It had been found that (i) with Discrete CousineTransform features, Jordan Element Network provided the best results followed by (ii) Discrete Wavelet Transform features that help in the diagnostic process in medical care.