Sperm recognition and multi-target trajectory tracking method

Abstract

The invention discloses a sperm recognition and multi-target trajectory tracking method, which comprises the following steps of: selecting a target sperm from a first frame of sperm image, acquiring coordinates of the target sperm, and taking the coordinates as starting coordinate points of a trajectory chain of the target sperm; calculating a prediction coordinate point of the target sperm in theith frame of sperm image by using a trajectory prediction algorithm; setting a search area according to the predicted coordinate points in the ith frame of sperm image for searching, if a plurality of sperms are searched in the area, performing association matching by combining multiple principles with a matching algorithm to obtain an optimal matching sperm, and adding the coordinate points of the optimal matching sperm into the trajectory chain of the target sperm; and if no sperm is searched in the area, adding the predicted coordinate point into the trajectory chain of the target sperm. According to the invention, the tracking efficiency of sperm motion trails is improved, the method is suitable for the occasion of temporary loss of sperms, the matching precision is improved, and themismatch rate is reduced.

Description

technical field [0001] The invention relates to the technical field of sperm sample computer analysis, in particular to a sperm identification and multi-target trajectory tracking method. Background technique [0002] The computer-aided semen analysis system can overcome the disadvantages of subjectivity, inefficiency, and error in the manual semen detection method to a certain extent, and can detect subtle morphological features and movement features that are not easy to be observed by the human eye. Image sequences can obtain more objective, precise and comprehensive sperm analysis results. [0003] In the process of analyzing sperm image sequences, it is necessary to perform inter-frame target tracking on multiple sperm, search for the position of the sperm in the next frame in the image sequence and correlate them, so as to obtain the trajectory of the sperm, especially in the process of multi-target matching In , the quality of matching results directly affects the cal...

AI Technical Summary

Problems solved by technology

[0005] 1. Search directly in a fixed large range in the image, there are many candidate targets, and the tracking efficiency is low

[0006] 2. If the corresponding matching target cannot be found in the next frame, the search will be terminated and no longer tracked

However, due to factors such as the similar appearance of sperm and irregular movement, the method based on this matching principle has a high mismatch rate, especially when the sperm density is high, the mismatch rate is even more unacceptable, and the calculated motion parameters are also difficult. measure the quality of sperm

Images