Edge detection of heterogeneity in transmission images based on frame accumulation and multiband information fusion

Abstract

In optical transmission imaging, even a very thin layer of a scattering material can appear opaque make the image blurry, which restricts the positioning of heterogeneity in transmission images. Aiming at the problem, an exploratory method based on frame accumulation and multiband information fusion was proposed and proven, which can be used to detect the edge of heterogeneity in transmission images. Treating the trichromatic image as an example, the following experiment was designed: Firstly, the transmission video of phantom is collected by camera. Then, single-band images are decoded at “R∖G∖B” and further accumulated by frame accumulation technology to obtain the image with high grayscale level, respectively. Lastly, the information fusion of the three single-band images is realized through neural network, and further the edge could be detected by the established model. The experimental result shows that the heterogeneity edge can be detected more accurately by this method. Therefore, the validity of this method for detecting heterogeneity in transmission images is demonstrated. It is expected to provide an idea for the future clinical diagnosis of breast tumor by transmission method.