Improvised Method Of Tumor Segmentation Using Saliency

Abstract- In recent years the image processing techniques are used commonly in various medical areas for improving earlier detection and treatment stages, in which the time span or elapse is very important to discover the disease in the patient as possible as fast, especially in many tumours such as the lung cancer, breast cancer. This system generally first segments the area of interest (lung) and then analyses the separately obtained area for nodule detection in order to examine the disease. Even with several lung tumour segmentations have been presented, enhancing tumour segmentation methods are still interesting because lung tumour CT images has some complex characteristics, such as large difference in tumour appearance and uncertain tumour boundaries. To address this problem, tumour segmentation method for CT Images which separates non-enhancing lung tumours from healthy tissues has been carried out by clustering method. The proposed method uses pre-processing technique that remove unwanted artifacts using median and wiener filters. Initially, the segmentation of the CT images has been carried out by using K- Means clustering method. To the clustered result, EK-Mean clustering is applied . Further the features like entrpy, Contrast, Correlation,Homogenity and the area are extracted from the tumorous part of Fuzzy Ek- Means segmented Image. For feature extraction, statistic method called Gray Level Co-occurrence Matrix (GLCM). Classification is done by using the

Mei wang, Xiao-Wei Wu, Hsiung-Cheng Lin, Jian-ping wang(2012) discussed the idea of following. Today medical images are widely used. To achieve a clear and noise free images Image-segmentation is essential point. In this paper a new image segmentation method that is proportion of foreground to background is used also called PFB. Using this method on human brain image we experiment the result that the number of iteration steps for the threshold T is reduced. Using this method over segmentation and discontinuities also can be avoided. Finally proposed a new method using both Thresholding image segmentation and edge detection segmentation algorithm based on the Proportion of Foreground to Background. After this the basic principal is define in this paper. Proportion of Foreground to Background Algorithm combine the traditional iterative algorithm.

Preprocessing method involves series of operation to enhance and make it suitable for segmentation. Main function of preprocessing is removal of noise that is generated during image generation. Filters like min-max filter, Mean filter, Gaussian filter etc. may be used to remove noise. Binarization process is used to convert grayscale image into black and white image. To enhance the visibility and structural information Binary morphological operation is used. This involves opening, closing, thinning, hole filling etc. The captured image may not be perfectly aligned so, slant angle correction is performed. Input image may be resized according to the need of

SIFT~\citep{SIFT} provides a set of features of an object that are are robust against object scaling and rotations. The SIFT algorithm consists of four main steps, which are discussed in the following subsections.

Canthus is initially defined and treated as a corner (interest point). Interest point detectors like Harris and SUSAN are being studied and discussed in this report. These corner detectors are evaluated to find which one suits best for this application.

The regions of an image containing different intensity values can be modelled as a finite Gaussian distribution. All the pixels in the same image belongs to same class and energy function is given by

Abstract— Photographing the changes in internal breast structure due to formation of masses and microcalcification for detection of Breast Cancer is known as Mammogram, which are low dose x-ray images. These images play a very significant role in early detection of breast cancer. Usually in pattern recognition texture analysis is used for classification based on content of image or in image segmentation based on variation of intensities of gray scale levels or colours. Similarly texture analysis can also be used to identify masses and microcalcification in mammograms. However Grey Level Co-occurrence Matrices (GLCM) technique introduced by Haralick was initially used in study of remote sensing images. Up till now in breast cancer detection only first and second order GLCM features were mostly used, to the best of our knowledge there is no evidence of use of higher order GLCM features for detection of malignant masses in breast tissue images. In this paper we attempted up to 7th order and observed the results by analyzing the effects of higher order features in recognition of malignancy in

Machine vision is a high-speed developing field in AI, and image identification is becoming more and more important with higher and higher requires of AI. As a big part of image identification, it is abviously necessary to develop better solution in image segmentation, thus machine could identity objects easier. The division technique for pictures based on Markov Random Field (MRF) demonstrate ready to combine the contextual information from label pictures and statistical properties of pictures to be segmented.

Classification between the objects is easy task for humans but it has proved to be a complex problem for machines. The raise of high-capacity computers, the availability of high quality and low-priced video cameras, and the increasing need for automatic video analysis has generated an interest in object classification algorithms. A simple classification system consists of a camera fixed high above the interested zone, where images are captured and consequently processed. Classification includes image sensors, image preprocessing, object detection, object segmentation, feature extraction and object classification. Classification system consists of database that contains predefined patterns that compares with detected object to classify in to proper category. Image classification is an important and challenging task in various application domains, including biomedical imaging, biometry, video surveillance, vehicle navigation, industrial visual inspection, robot navigation, and remote sensing.

Many papers address image segmentation; however, this section only reviews papers that have tried to improve on FCM-based image segmentation. In its original design, the FCM algorithm assigns a membership value to each pixel and each image cluster. For image I with a grayscale range of for the ith pixel ( in k-dimensional space , the cluster centers are defined as . Parameter C denotes a positive value ( ) and is the membership value assigned to the ith pixel in the jth cluster. The respective objective function in FCM is defined as [23]:

Over several past years, contrast image enhancement has generated across many applications like robot sensing, electronic products, fault detection, medical image analysis, etc. Thus, increasing in popularity of contrast enhancement of images has forces researchers to study their enhancement techniques and their effectiveness for the interpretability or perception of human viewers. Contrast enhancement is a

Over several past years, contrast image enhancement has generated across many applications like robot sensing, electronic products, fault detection, medical image analysis, etc. Thus, increasing in popularity of contrast enhancement of images has forces researchers to study their enhancement techniques and their effectiveness for the interpretability or perception of human viewers. Contrast enhancement is a vital part of various fields, such as X-ray image analysis, biomedical image analysis, machine vision where pixel

Feature plays a very important role in the area of image processing. Different feature extraction

This representation of Haar-Like Feature was used to find and select features according to pixel intensities and not by pixels values. Haar-like feature can be calculated using the scalar product between the input image and some Haar-like templates [14].

In pattern recognition and image processing, feature extraction is a special form of dimensionality reduction. The main goal of feature extraction is to obtain the most relevant information from the original data and represent that information in a lower dimensional space. Effective feature extraction from various intensity or color in images have been an important topic