Design And Realization Of Sperm Morphology Analysis System Based On OpenCV2

Sperm morphology analysis plays an important role in the evaluation of sperm fertilizing capacity, sperm pathology analysis, and other aspects of assisted reproduction and contraception. Currently, sperm morphology analysis is mainly sperm micrograph analysis after staining, and classification according to WHO criteria. Since this work is mainly done by the artificial, the result is subjective and ineffective. With the development of computer technology, digital image processing technology and computer vision, computer-assisted sperm morphology analysis systems (CASMA) has arisen. Detection with CASMA is quickly consistency, and can gives numerical results, which would help medical practitioners make decisions and treatment options. So, creating an accurate and effective system has high value in clinical diagnosis and scientific.This work aims to provide sperm morphology accurate data for clinical diagnosis of reference and scientific basis. Sperm morphology and related parameters including abnormal sperm head, acrosome, mid-piece and other parts of the sperm. The world health organization’s latest "Human semen examination and processing laboratory manual" provides the reference value and limit threshold for the semen analysis system. In this paper, the work was performed as described in the manual.Combined with domestic and foreign research and product, we use OpenCV match VS development environment to analysis sperm morphology, and extract sperm morphological features to classify them.For the convenience of processing, preprocess the original image. The processed image is made to reach the similar level as much as possible. Preprocessing includes filtering, morphological operation, Binarization operation and so on. Then extract a region of interest from pre-processed image, and thresholding ROI with the threshold value calculated by OTSU. Filter out impurities according to their areas, and calculate the overall outline. Use morphology operation to remove the mid-piece. Then according to the acrosome initial contour, using the GVF Snake model implementation for segmenting the acrosome, and repeatedly adjust its parameters in order to obtain best segmentation results finally. Further to calculate parameters such as area, angle of each part, and according to the standard analysis the parameters. We also designed the sperm morphology detection system interface, complete the image read, call dll files, parameter returns and result display, etc.Finally compare the system test results with manual test results and the results is satisfactory. System to further improve after can realize fully automated sperm morphology test, improving efficiency, unifying inspection standard. In this paper, the research content has high reference value in image analysis and clinical study.