Blood Cell Detection Using Thresholding Estimation Based Watershed Transformation with Sobel Filter in Frequency Domain

Blood cells detection in microscopic image provides the information concerning the health of patient. The analysis of blood cells using image processing reduces the manual disease detection error and also the time period. A new thresholding estimation algorithm has been proposed with watershed transforming Sobel filter in frequency domain for detection of different cells in microscopic image. The proposed algorithm performs edge detection using Sobel filter in frequency domain. The present study of Sobel filter uses specific window size scheme to remove noises and detect the fine edges. Consequently, thresholding estimation based watershed transformation is used on the specific window size Sobel filter to increase the intensity of edges with strong contrast. Thus this effective detection algorithm is helpful to identifying and counting the different cells. In this study, proposed algorithm has used 30 numbers of blood microscopic images as test images and obtained higher accuracy results of around 93%. Experimentally, the proposed algorithm yields better structure similarity index measure, compared with the other state-of-art detection method viz. Canny, Sobel and Laplacian operator.