Two Filters Based on Simple Functions for Extracting Profiles from Images

Graphic contour extraction is a significant issue in pattern recognition and vision processing. However, classical filters such as Sobel filters are difficult to extract the complete shapes in terms of contour extraction. In this work, we designed two filters via logarithmic, sine, exponential, and cosine functions. These filters can extract the outline of the image completely. Moreover, a low-light image acquired from a fluorescence microscope was used for contour extraction. These filters could be used for extracting some of the edge contours for images with weak fluorescence and underwater images, where colouring impact was applied to the outputting images for generating coloured images. Meanwhile, the successful extraction of crystal contours under the fluorescence microscope via these filters proved their feasibility for microscopic image processing.