Mycobacterium tuberculosis detection method and device, terminal equipment and storable medium

The invention provides a mycobacterium tuberculosis detection method, which comprises the following steps: determining a target feature extraction backbone network Backbone of a two-layer structure of mycobacterium tuberculosis, extracting target feature information from an acquired mycobacterium tuberculosis microscopic image, and generating a C1 layer target feature map and a C2 layer target feature map; generating a plurality of anchor frames on the C2 layer target feature map, and extracting a plurality of regions of interest from the plurality of anchor frames; according to the region of interest and the extracted target feature information, constructing a mycobacterium tuberculosis detection model of a lightweight neural network; and inputting a mycobacterium tuberculosis microscopic image to be detected into the mycobacterium tuberculosis detection model of the lightweight neural network, and outputting a corresponding mycobacterium tuberculosis detection result. According to the method, the detection