Lesion Area Detection (LAD) using Superpixel Segmentation

Nowadays, in biomedical field, imaging techniques are used to discover the abnormalities in the particular region. Lesion is an area in an organ or tissue, which has suffered from damage through injury or disease, such as a wound, ulcer, abscess, or tumor. Lesion area of the tissue is detected using source images. Wireless capsule endoscopy image of the gastrointestinal tract is used to find the lesion. Preprocessing of image is done using denoising by multidimensional filter. Denoised image is edge detected during which the edge points in an image differentiation is highlighted by Sobel edge detection. The value of each pixel in the edge detected image is evaluated based on the comparison of corresponding pixel in the input image with its neighboring pixels by the morphological operation. Erosion and dilation operation erodes and adds pixels respectively to the edge detected image, so that the value of the output pixel has the maximum pixel value. During segmentation, the pixels are grouped using superpixel algorithm to reduce the complexity while maintaining high diagnostic accuracy. As a final point, the Correlation coefficient test finds the spoiled area between healthy tissue and abnormal tissue by color encoding. This intends to identify the suspected abnormal areas (lesion) from the source images.