Automatic sperm morphology analysis method based on multi-scale feature fusion

Abstract

The invention belongs to the technical field of sperm morphology detection, and particularly relates to an automatic sperm morphology analysis method based on multi-scale feature fusion, which comprises the following steps: scanning a sperm sample image through a microscope automatic scanning platform, performing normalization processing on the obtained sperm sample image, extracting features of sperms on multiple scales by using an SD Net network model, fusing the features, outputting coordinates and bounding box information of the sperms on the sperm sample image, and detecting sperm images existing in the sperm sample image; and carrying out feature extraction and regional feature fusion on the detected sperm image by using an SR Net network model, and finally completing the recognition of abnormal sperm cells. Compared with a general deep learning method, the invention has the advantages that by fusing the features on multiple scales, the method has better detection capability on sperms with smaller sizes, the detection accuracy is higher, the generalization is stronger, and a very good detection effect can also be obtained under the condition of larger sample image difference.

Description

technical field [0001] The invention belongs to the technical field of sperm morphology detection, in particular to an automatic sperm morphology analysis method based on multi-scale feature fusion. Background technique [0002] The results of sperm morphology testing are one of the important criteria for judging sperm quality. Studies have shown that there is a great relationship between sperm morphology and sperm motility, and men with teratozoospermia (less than 4% of normal sperm count) may have a higher Sperm DNA damage risk, therefore, sperm morphology detection is an important method to evaluate male fertility and fertility quality. [0003] Traditional sperm morphology detection is done manually by experts, and the results are highly subjective and time-consuming. Automated sperm analysis can minimize human factors, but traditional sperm analysis is based on wavelet image enhancement technology and gradient analysis to find the direction of sperm, and then perform s...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the deficiencies of the prior art, and propose a sperm morphology analysis method based on multi-scale feature fusion, which solves the problem of poor robustness of the traditional sperm morphology analysis method, and in the High accuracy can be achieved even when the mass difference is large, the background impurities are large, and the target is small

Images