Sperm morphology by image analysis compared with subjective assessment.

Human sperm heads were detected and tracked using semi-automatic image analysis. Six measurements of size and shape were computed from two samples of semen from each of 26 men. Subjective assessments of abnormal morphological forms were made on the same samples by two independent assessors. A formula for prediction of percentage abnormality was calculated from the results. The subjective assessment of abnormality was related primarily to the number of elongated heads and secondly to the number of small heads. Computed image processing provides an objective method for measurement of the morphological features of sperm.