Spore and spore identification method and system based on residual deep network, and medium

Abstract

The invention provides a spore identification method and system based on a residual error deep network and a medium, and the method comprises the steps: 1, segmenting a collected original picture through employing a morphological method, screening a target picture needing to be identified, forming a sample set, and marking the sample set; 2, performing data amplification on the labeled sample set to form a final training data set; 3, constructing a deep residual network on the caffe framework, setting hyper-parameters, and training by using the data set to obtain a deep residual network training model; and 4, performing morphological segmentation on an original picture collected in real time, performing identification scoring by using the deep residual network training model with the highest comprehensive identification rate after the test, and taking a result greater than a confidence threshold as a final identification result and outputting the final identification result. According to the method, the spores and the bud spores are identified by adopting the residual deep network structure, so that the problems that morphological identification needs a large number of parameters, the efficiency is low and the identification rate is low are solved.

Description

technical field [0001] The present invention relates to the technical field of deep learning, in particular to a method, system and medium for identifying spores and spores based on a residual deep network. Background technique [0002] Leucorrhea is the secretion of female vagina, and the microecological flora detection of leucorrhea is an important condition for judging whether the female reproductive system is healthy. Among them, fungal infection is already a major disease that seriously affects people's health in my country. Spores and blastospores are early stage symptoms of fungal infection. The early detection of the disease is self-evident for hospitals and patients. . However, at present, the ability of medical units to detect fungal infections needs to be improved urgently, especially the urgent need to adopt new scientific and technological methods to detect pathogenic bacteria accurately and efficiently. [0003] The main techniques for detection include direct...

AI Technical Summary

Problems solved by technology

However, this method requires the operator to have rich inspection experience, and it is prone to false positives and false negatives for test specimens with a small amount of bacteria and complex background components.

Compared with the existing automatic biochemical detection equipment, artificial fungal microscopy has the disadvantages of high labor intensity and high degree of subjective judgment of the results.

Images