Light refraction law based image processing

The subject ?image processing? has become such a subject that has a growing importance and application area nowadays. Today we encounter with the applications of image processing in many fields like medicine, security, industry, earth sciences, history, art and so on. In this study a new approach for image processing has been developed. The new approach is based on the law of light refraction. The improved algorithm basically focuses to find the similarity relation values of pixels. In the calculation of similarity ratios of each pixel, the light refraction has emerged which is a physical event. After the light has been refracted in a transparent sheet, it gets through from the entrance direction. However as a result of the difference between incident angle and refraction angle of light, there exists a parallel shift. This shift amount has been used for measuring similarity between pixels. While edge detection is being done, a similarity relation matrix is set for each pixel in the image and the biggest similarity value in this matrix is chosen and is located in the image. Hence the edge map of the image is obtained. Also in the segmentation operation the similarity values of each pixel is found and compared with the determined threshold value. As a result of this comparison, the current pixel is labeled and it is included in a region consisting similar pixels. The key point here is a pixel included by a region must take a positive label value and cannot be included in any other region. The label value of an unprocessed pixel must be negative.The method used in this thesis is advantageous in terms of having not complex calculations. The proposed method has been tested on the standard images. Satisfactory results have been attained by using this method. This method can be used in the further studies about image filtering.