ABSTRACT
Early computer diagnosis is necessary for the quantitative analysis in computer vision-based cancer image analysis, which requires segmentation as an essential stage. Due to the shape, size, and location of pancreas in abdominal Computed Tomography (CT) images, the segmentation process of these CT images is considered a cumbersome task. Besides, variations in appearances, consistency/inconsistency, and features have a great impact on the classification. Hence, extracting visual clues related to the pixel level labeling is considered the major task in diagnosis applications. High anatomical variations in the pancreas affect the visual clues-based segmentation methods and limit the accuracy as compared to other organ segmentation from abdominal images. This chapter proposes an integrated framework of Gaussian with the level set formulation to avoid the probable mask creation. Initially, the proposed work employs the Laplacian formula to filter the noisy regions from the images and the second-order formulation to sharpen the edges of images. Then, the integral framework of Gaussian with the level set formulations is applied to segment the pancreas from these images. Then, the N-ternary-based methodology is used to extract different patterns from the images. Finally, the datasets are trained using Convolution Neural Network (CNN), and the resulting N-ternary pattern modules depict the labels for classification. The comparative analysis between the proposed CNN-based learning model with the various existing methods and evaluation of the performance parameters ensures their effectiveness in early detection models.
