ABSTRACT Recently, image process techniques square measure wide employed in many medical areas for image improvement in earlier detection and treatment stages, where the time issue is incredibly necessary to find the abnormality problems like stones in gallbladder, carcinoma, breast cancer, lung cancer etc. Image quality and accuracy is that the core factors of this analysis. There are 3 stages to be done for detecting stones in gallbladder. The first stage is Image enhancement in which we use the median and mean filters to reduce noise in the image. Following the segmentation principles with Thresholding approach and Watershed segmentation approach. The last stage is Features extraction which will be also found by Binarization approach and Masking approach. A normal comparision is made. In this research, the main aim is to detect stone in gallbladder and find its size. INTRODUCTIONGallstones are solid stones formed from bile precipitates. A CT scan image of gallbladder is taken to detect the presence of stones in it. This Image has to be processed in the following 3 stages.1)Image enhancement2)Image segmentation 3)Features extractionThe main aim of image enhancement is to improve the perception of information included in the image for human viewers. Segmentation divides the image into its constituent regions. Image features Extraction stage is significant stage that uses algorithms and techniques to isolate and detect various desired portions or shapes (features) of a given image. LITERATURE SURVEY For studying the concepts of image enhancement, I have gone through some papers in medical image enhancement. As we know that medical images are not captured clearly due to some noise, blur and many more. But it is difficult to deal with all the problems together. Here I am discussing about some previous papers. Vinita Malik et al (2006), proposed a comparative study of image denoising using mean and median filtering on digital images. In this study, they established that the most of the median filtering methods used in more than one frame work in order to achieve an improved impulse noise cancellation. Impulse noise can be mathematically explained by several equations. In general, impulse noise can appear as bright and dark spots on the image. These spots have high contrast when compared to surrounding spaces. Prof. S. T. Khandarel et al (2014) surveyed a paper on Image Segmentation with Thresholding. The main scope of work is development of an most effective approach for the color image segmentation, where we will dicuss the concept of thresholding by referring the color image statistical data. To achieve the aim of the proposed work, we will make advancement in the thresholding mechanism. To achieve thresholding, global and local global analysis, and assessment of the color image data will be discussed. Main aim of the work is to enhance a segmentation approach for color images based on colour models and an automatic thresholding. To achieve the main focus of the work is automatic thresholding and Quad-Tree decomposition by constituting image into different color models. In management of blurry objects, which have unclear edges from an out-of-focus , the estimated method can also be segmenting the necessary objects. The proposed method could take the less computational time to find strong and high-quality segmentation performance than the traditional ones. Therefore, the planned method can be widely and successfully working in various segmentation applications. R. Kiruthikaa Implemented Watershed Based Image Segmentation Algorithm Using FPGA Processor. In this paper, the watershed algorithm is used as a method in resolving the image segmentation problem. However, many variants of the watershed algorithm all are not equally well suited for hardware implementation. Watershed algorithm are suited and different algorithms based on connected components is selected for the implementation as it displays complexity good segmentation quality, least computational and can be implemented in FPGA. The watershed segmentation have disadvantage of over segmentation. The markers based watershed segmentation transformation is discussed in this project. This watershed approach reduces over segmentation and noise. It has rationalized memory access compared to all other watershed based image segmentation algorithms. The main focus of this project is to execute image segmentation algorithm in a FPGA that requires minimum low execution time, hardware resources, and is suitable for use in real time applications. It has been shown that good segmentation results can be attained. The proposed algorithm is a powerful and efficient tool for image segmentation, even for overlapping. These implementations, when combined with the process of mark selection, permit an powerful approach for image segmentation. The result of this combined procedure is however, depends upon the process of mark selection. Sudeep Thepade et al (2014) presented a Novel Feature Extraction Technique Using Binarization of Bit Planes for Content Based Image Classification. The operation was followed by binarization of important bit planes for feature vector extraction. Binarization approach calculated the threshold value to recognize the object of interest from its background. This method has been compared with regard to quantity for the techniques proposed by Kekre et al. in and by Thepade et al. in and four other extensively used image binarization techniques proposed by Bernsen, Sauvola, Niblack, Pietikäinen and Otsu. Mean square error (MSE) method was followed for categorization performance estimation of the Binarization technique with respect to the present techniques for feature extraction. The planned technique has implemented feature extraction by binarization of important bit planes with Niblack’s local threshold selection method. Binarization of important bit planes was done with mean threshold method. Another present used technique of feature extraction has also used threshold for binarization. Relative analysis of misclassification rate (MR) and F1 score of the planned technique of feature extraction has well performed the feature extraction method. Mokhled S. AL-TARAWNEH(2012) proposed Lung Cancer Detection Using Image Processing Techniques an image enhancement technique is developed for earlier disease detection and treatment stages. The time factor was taken in account to find the abnormality issues in medical images. Image accuracy and quality are the core factors of this research. Image quality analysis and enhancement stage where there were adopted on low pre-processing techniques based on Gabor filter within Gaussian rules. The masking technique is well organized for segmentation principles to be a region of interest for feature extraction obtaining. This technique gives very good results comparing with other existing techniques. Depending on general features, comparison is made. Pixels percentages is the main feature for accurate images comparison. METHODOLOGYDenoising step has been done using mean filter. The results found are Peak Signal to Noise Ratio(PSNR), Mean Squared Error(MSE).MSE measures the average of squares of error. MSE = sum(sum((I-I1).^2))/(M*N) where I and I1 are input image and denoised image respectively.PSNR is the ratio between the maximum possible power of the signal and the power of corrupting noise. PSNR = 10*log10(256*256/MSE) (a) Original image (b)Enhance image Fig.1 (a) Original image (b)Enhance imageFig.2 (a) Original image (b)Enhance image Fig.3 RESULTS ANALYSIS Mean Filter Parameters Fig. 1 Fig. 2 Fig. 3 MSE4.414.334.57PSNR41.724741.795241.569 Median Filter Parameters Fig. 1 Fig. 2 Fig. 3 MSE1.01401.16321.512PSNR48.104647.508147.1917 CONCLUSION A comparision of mean and median filter for denoising is made by finding the parameters MSE and PSNR. The quality of image depends on PSNR value. As the denoised image obtained by median filter have high PSNR value than image obtained by mean filter. Median filter is more efficient than mean filter.