Automatic analyzing method and system for vaginal microecological morphology

Abstract

The invention aims at providing an automatic analyzing method and system for vaginal microecological morphology. A low-resolution smear image obtained through conventional equipment is subjected to super-resolution amplification rebuilding to be clearer, a foundation is laid for follow-up classification, then, a nerual network technology is applied for judging and recognizing the rebuilt high-resolution image, and the variety and density of microorganisms are further output, so that the purpose of carrying out automatic analysis and computer assisted recognition on vaginal microecological morphology is achieved, and objective and accurate clinic evaluation is carried out on a vaginal microecological system.

Description

technical field [0001] The invention relates to a method and a system for micro-ecological morphological analysis, in particular to a method and system for automatic analysis and computer-aided identification of vaginal micro-ecological morphological images. Background technique [0002] Female reproductive tract infection and sexually transmitted infection are common diseases among women, and they are also global social and public health problems. They have the characteristics of high incidence, high recurrence rate, and wide epidemic range. Infections are the source of many diseases, and if genital tract infections are not promptly diagnosed and properly treated, they can lead to HIV infection and an increased risk of developing cervical cancer, infertility, ectopic pregnancy, miscarriage, stillbirth, stillbirth, premature birth, Complications such as congenital infection, neonatal infection, and chronic abdominal pain affect the health of two generations. [0003] The es...

AI Technical Summary

Problems solved by technology

[0005] Existing computer image recognition techniques, such as the Wavelet PCA method (i.e. wavelet principal component analysis based on wavelet transform, Wavelet Principal Component Analysis), can extract image features relatively quickly, but this method is relatively simple for image processing, only relying on Priori knowledge of bacterial characteristics can be used to classify, but the recognition ability for diverse bacterial smear images is extremely limited

However, image recognition using artificial intelligence methods has a high accuracy rate, but the calculation is complex and often requires large-scale parallel computing capabilities to complete, and the image resolution obtained by existing equipment for obtaining patient smears is relatively low. Low, artificial intelligence operations such as neural network learning cannot be effectively performed, and the equipment for obtaining high-resolution images is expensive and takes a long time to process, so it is not suitable for clinical smear examinations for a large number of patients

Images