Zhang, et.al,” An improved watershed algorithm for color image segmentation,”, 2012.The author18 have demonstrated the adaptive marker extraction -based watershed algorithm is used to overcome the over-segmentation problem.
It also analyzes the disadvantages of the classical watershed segmentation and presents a marker extraction based on adaptive color image segmentation algorithm to improve the watershed, the conventional marker for the lack of extraction methods, many consider the minimum characteristics of properties, and set the adaptive threshold. This method can efficiently reduce over-segmentation with scarcely computational complexity increase. It has improved anti-noise performance and edge-location ability. The improved watershed algorithm includes three steps:1. Computing the vector gradient of the color image.2. Extracted the low frequency component of the vector gradient image.
3. The extracted marker as the minimum of original gradient image, and then watershed transformation is used to process the gradient image with minimum marker. 11. T.Roman Singh, et.al, “A New Local Adaptive Thresholding Technique in Binarization”, 2011.The author19 used an efficient way to determine local threshold,. Binarization can be speed up considerably as a result of using the integral sum image(Integral sum image is determined as a prior process for determining the local sum .
If local sum is available, mean can be calculated with a simple arithmetic operation) to determine the local sum, where computational time does not depend on the window dimension. In other techniques like Sauvola’s and Niblack’s technique, local mean and standard deviation are required to determine the value of the threshold for each pixel. In this proposed technique, no local standard deviation is used and it requires to compute the local mean and mean deviation to determine the local threshold. Calculation of local mean deviation is straightforward by just subtracting the mean from the concerned pixel. Thus the proposed technique can binaries faster than others.
12. Y. Wangsheng, et.al, “Color Image Segmentation Based on Marked-Watershed and Region-Merger”, 2011.
In this paper22, A color image segmentation algorithm combined marker-based watershed and region merger was proposed to deal with over-segmentation. Firstly, it extracted minima-marker adaptively according to the information of local minima of gradient, and watershed the marked gradient image to get the pre-segmentation result. Then, it modified region distance according to real sense of region-similarity of human-vision, and defined an integration distance measurement considering color distance, color difference and edge information, which would lead the pre-segmentation regions to merge to a final segmentation result. It results in well suppress the over-segmentation, which owns a lower Local Consistency Error and are more accordant to the human-vision segmentation. 13.
N. Anh, et.al, “Morphological Gradient Applied to New Active Contour Model for Color Image Segmentation”, 2012.They 23 proposed a novel segmentation algorithm based on active contour models, to overcome the weakness. First, a morphological gradient -based edge detection is applied to an image, it helps to avoid losing color characteristics, compared with gray scale conversion. Second the edge map will be used as a clue to provide both good edge information and region information for an active contour without a re-initialization model.
As a result, proposed algorithm allows the contour to the initialized more flexibly, evolves the contour faster, and segments the boundary of subjects more precisely i color images.14. Ayesha Khalid Khan, et.al, “Gulistan Raja and Ahmad Khalil Khan, “Implementation of Marker based Watershed Image Segmentation on Magnetic Resonance Imaging,”Life Science Journal”, 2013. In this paper24 implemented marker based watershed segmentation technique is used by applying different detection operators on MRI test images. The image is converted into grey scale and then its gradient magnitude image is found by different detection operators.
The threshold image is taken to show the image details in binary representation. The binary image is processed using marker based watershed technique for segmentation where different morphological operations are performed. Image parameters of the segmented image with respect to the input image in the optimum range are computed.
The output shows the successful segmentation of different regions of MRI. 15. Nassir Salman, “Image Segmentation Based on Watershed and Edge Detection Techniques”,2006.In27,using multi-threshold which is important to eliminate false edges and thus obtain larger regions, the DIS map consists of all edge information about the input image even on the smooth regions, and a combination of K-means, watershed segmentation method was used to perform image segmentation and edge detection tasks. An initial segmentation based on K-means clustering technique. Starting from this, they have used two techniques; the first is watershed technique with new merging procedures based on mean intensity value to segment the image regions and to detect their boundaries. The second is edge strength technique to obtain an accurate edge maps of images without using watershed method. The problem of undesirable over segmentation is solved in this paper by these techniques, when used directly with raw data images.
Also, the edge maps obtained have no broken lines on entire image and the final edge detection result is one closed boundary per actual region in the image. DIS map are good techniques to perform image segmentation and edge detection tasks.16. Anju Bala, et.al, “Split & Merge: A Region Based Image Segmentation”, 2017.Split and merge algorithm is used a sequential segmentation algorithm.
The SM method based on different homogeneity criterion results in different results. Hence, the best homogeneous criterion for merging and splitting is difficult to find.29 The appropriate merge rule is used to produce final segmentation results.
This process is dividing into four phases split the image, merge similar sub regions and spatially adjacent regions and elimination of small regions. In this method image is considered as a piece at a time called tile. It is based on the pixel partition mode test. This method can easily detect slowly changing regions. Different split and merge segmentation based methods are used in this paper. Moreover, SM is an iterative algorithm which consumes more time and space and depends on the local properties and there is no simple way to add global properties of image.
17. Andrei C. Jalba, et.al, “Interactive Segmentation And Visualization of DTI Data Using A Hierachical Watershed Representation”,2015.The author proposed an interactive segmentation approach, based on a hierarchical representation of the input DT(Diffusion Tensor) image through a tree structure.
The tree is obtained by successively merging watershed regions, based on the morphological waterfall approach, hence the name watershed tree. Region merging is done according to a combined similarity and homogeneity criterion. Filters introduced that work on the proposed tree representation, and that enable region based attribute filtering of DTI data. Linked views between the visualizations of the simplified DT image and the tree enable a user to visually explore both data and tree at interactive rates.35 The coupling of filtering, semiautomatic segmentation by labeling nodes in the tree, and various interaction mechanisms support the segmentation task. Once the tree representation has been computed, it can be saved to external memory, for later usage. If multiple regions have to be segmented, or if a segmentation has to be changed/refined, the whole initialization and level-set propagation process has to be repeated.
Moreover, if the result after propagation does not meet certain requirements parameter adjustment has to be done, and the whole process has to be repeated.