Surrounds methods of image objects

In this thesis study, methods which are called active contour models (ACMs) in the literature and which enable any objects in the image to be surrounded are investigated. In the segmentation process, groups of pixels exhibiting a homogeneous color distribution in the image are grouped. ACMs are studied under the topic of segmentation in the region of image processing. The main purpose of the ACM is to discover the boundary line that separates the foreground object from the background area. When the ACMs in the literature are examined, it is seen that there are two classifications, edge-based and region-based. Area based ACMs are based on the idea that "the foreground object is in the middle of the view". Edge-based methods use a continuous approach to the original image function. Edge-based ACMs use image gradient-based edge detection to achieve the ultimate perimeter. Region-based methods are based on edge-based methods are better known. For this reason, the focus is on region-based methods in the thesis study. To mention briefly region based methods; The C-V method, an improved version of the Mumford-Shah model, can detect the boundary line surrounding the foreground object without the need for gradient use. The SBGFRLS method continuously updates a mask to cover the objects in the image, thereby trying to separate the foreground object from the background. ORACM is an enhanced version of the SBGFRLS method. ORACM is a region-based ACM that does not require parameters and requires less time. Fuzzy energy based active containment method (FACM) uses fuzzy logic theory to detect active boundary lines, unlike the conventional methods mentioned above. FACM uses global and local image information together. In the thesis, the FACM method is examined in detail. In this thesis study, region based ACM applications were performed using Matlab. The advantages and disadvantages of the methods tested using real images with different constructions have been extensively discussed. As a result, it has been found that FACM and ORACM methods accurately detect the active boundary line and produce close results. KEYWORDS: Active Contour Model, C-V, SBGFRLS, ORACM, FACM