 |
|
|
|
|
 |
 |
|
Computational Model for Sperm Cells Identification and Morphological Classification by Automatic Analysis of Digital Micrographs
|
|
|
|
|
|
| Organization: | Wolfram Research |
|
|
|
|
|
|
Wolfram Technology Conference 2015
|
|
|
|
|
|
Champaign, Illinois USA
|
|
|
|
|
|
The American Society for Reproductive Medicine states that the analysis of sperm cells is the first and most important step in the evaluation of male fertility, being their morphology one of the best discriminators of the fertilization potential. Nowadays, the methods used to evaluate sperm morphology are mostly manual and lack objectivity. This work proposes a computational model to identify sperm cells in digital micrographs and analyze whether head’s morphology is normal or not. The model had been developed using the Wolfram Mathematica environment in all the stages, which are the extraction and selection of the features that identify the head of the sperm cell and later on, their classification in normal or abnormal attending to their morphology. Finally, there is a comparative analysis with the World Health Organization’s standards and criteria for sperm cell morphology. The resulting model allowed the identification the 91.5% of the sperm cells that were contained in the provided images. The accuracy percentage of the head’s morphology classification was 77.6%. These results are successful when compared to the ones obtained in previous studies.
|
|
|
|
|
|
|
|
|
 |
|
|
| 1445539676.nb (227.2 KB) - Mathematica Notebook |
|
|
For the newest resources, visit Wolfram Repositories and Archives »
