A cell detection and recognition system and method

Abstract

This invention relates to the field of image analysis technology, and more particularly to a cell detection and recognition system and method. The method includes the following steps: acquiring an image sequence of a first imaging region; preprocessing the image sequence; inputting the image sequence into an instance segmentation network to segment it into a candidate particle set; extracting image features from each candidate particle in the candidate particle set; inputting the image features into an image classification model to output the overall count information of the first imaging region, including the initial number of bacteria, the initial number of fragments, and the total number of particles; acquiring an image of a second imaging region; segmenting the second imaging region image using a semantic segmentation network to obtain a channel region mask; performing instance segmentation and counting within the channel region mask to obtain sample count information; and constructing a state vector for sampling decision. This invention, based on image analysis technology, improves the accuracy, statistical stability, and automation level of bacterial identification while reducing the cost of spectral acquisition.

Images