الفهرس | Only 14 pages are availabe for public view |
Abstract Image processing [1, 2] is the technique of converting an image into its digital format and performing manipulations on it to enhance its quality or to extract some valuable information from it. Many different techniques were introduced to perform different tasks on the image. Recently many new approaches are introduced which are inspired from nature (Bio-inspired). They follow the principle of charles Darwin of “Survival of Fittest”. The solutions which are best among vast pool of solutions are only forwarded to next generation or next iteration step and rest are discarded. Neural Network are shapes of Bio-inspired techniques. [1, 3, 19]. Traditional methods and Bio-inspired techniques works in the area of image processing, specially convolution neural network which can be applied in the field of enhancement techniques of images for example: image segmentation using edge detection [8, 9,21], content based image retrieval [10] and image restoring [11, 12]. In some or the other aspect these algorithms have enhanced the performance of each image processing application, these techniques are gaining more interest day by day due to their effective performance in every research domain [13]. In this thesis, we study some methods for solving this ill-posed problem which is the segmentation of images. In addition, introduce a new segmentation method for the breast and evaluate the effectiveness by solving the problem of detecting breast cancer from thermal images, comparative studies are presented. Parts of this thesis were published and presented in 9th Cairo International Biomedical Engineering Conference (CIBEC). |