Impact of transfer learning for human sperm segmentation using deep learning.

BACKGROUND AND OBJECTIVE: Infertility affects approximately one in ten couples, and almost half of the infertility cases are due to the malefactor. To diagnose infertility and determine future treatment, a semen analysis is performed. Evaluation of sperm morphology is one of several steps in semen analysis, in which the shape and size of sperm parts are examined. The laboratories dedicated to this use traditional methods susceptible to errors. An alternative to replace the poor visual ability to assess sperm size and shape is to analyze sperm morphology with a computer's help. However, since the automatic sperm classification rates do not show an acceptable precision rate for use in the clinical setting, it is considered an exciting approach to focus efforts on improving the precision in sperm segmentation to extract the contour sperm before classification. This work aims to assess the utility of two image segmentation deep learning models for segmenting human sperm heads, acrosome, and nucleus. METHODS: In this work, we evaluate the use of two well-known deep learning architectures (U-Net and Mask-RCNN) to segment parts of human sperm cells using data augmentation, cross-validation, hyperparameter tuning, and transfer learning. The experimental results are carried out using SCIAN-SpermSegGS, a public dataset with more than two hundred manually segmented sperm cells and widely used to validate segmentation methods of human sperm parts. RESULTS: Experimental evaluation shows that U-net with transfer learning achieves up to 95% overlapping against hand-segmented masks for sperm head (0.96), acrosome (0.94), and nucleus (0.95), using Dice coefficient as the evaluation metric. These results outperform state-of-the-art sperm parts segmentation methods. CONCLUSIONS: The impact of transfer learning is substantial, significantly improving the results of state-of-the-art methods with a higher Dice coefficient, less dispersion, and fewer cases where the model failed to segment sperm parts. These results represent a promising advance in the ultimate goal of performing computer-assisted morphological sperm analysis.

Automatic classification of human sperm head morphology.

BACKGROUND AND OBJECTIVE: Infertility is a problem that affects up to 15% of couples worldwide with emotional and physiological implications and semen analysis is the first step in the evaluation of an infertile couple. Indeed the morphology of human sperm cells is considered to be a clinical tool dedicated to the fertility prognosis and serves, mainly, for making decisions regarding the options of assisted reproduction technologies. Therefore, a complete analysis of not only normal sperm but also abnormal sperm turns out to be critical in this context. This paper sets out to develop, implement and calibrate a novel methodology to characterize and classify sperm heads towards morphological sperm analysis. Our work is aimed at focusing on a depth analysis of abnormal sperm heads for fertility diagnosis, prognosis, reproductive toxicology, basic research or public health studies. METHODS: We introduce a morphological characterization for human sperm heads based on shape measures. We also present a pipeline for sperm head classification, according to the last Laboratory Manual for the Examination and Processing of Human Semen of the World Health Organization (WHO). In this sense, we propose a two-stage classification scheme that permits to classify sperm heads among five different classes (one class for normal sperm heads and four classes for abnormal sperm heads) combining an ensemble strategy for feature selection and a cascade approach with several support vector machines dedicated to the verification of each class. We use Fisher's exact test to demonstrate that there is no statistically significant differences between our results and those achieved by domain experts. RESULTS: Experimental evaluation shows that our two-stage classification scheme outperforms some state-of-the-art monolithic classifiers, exhibiting 58% of average accuracy. More interestingly, on the subset of data for which there is a total agreement between experts for the label of the samples, our system is able to provide 73% of average classification accuracy. CONCLUSIONS: We show that our system behaves like a human expert; therefore it can be used as a supplementary source for labeling new unknown data. However, as sperm head classification is still a challenging issue due to the uncertainty on the class label of each sperm head, with the consequent high degree of variability among domain experts, we conclude that there are still opportunities for further improvement in designing a more accurate system by investigating other feature extraction methods and classification schemes.

Gold-standard for computer-assisted morphological sperm analysis.

BACKGROUND AND OBJECTIVE: Published algorithms for classification of human sperm heads are based on relatively small image databases that are not open to the public, and thus no direct comparison is available for competing methods. We describe a gold-standard for morphological sperm analysis (SCIAN-MorphoSpermGS), a dataset of sperm head images with expert-classification labels in one of the following classes: normal, tapered, pyriform, small or amorphous. This gold-standard is for evaluating and comparing known techniques and future improvements to present approaches for classification of human sperm heads for semen analysis. Although this paper does not provide a computational tool for morphological sperm analysis, we present a set of experiments for comparing sperm head description and classification common techniques. This classification base-line is aimed to be used as a reference for future improvements to present approaches for human sperm head classification. METHODS: The gold-standard provides a label for each sperm head, which is achieved by majority voting among experts. The classification base-line compares four supervised learning methods (1- Nearest Neighbor, naive Bayes, decision trees and Support Vector Machine (SVM)) and three shape-based descriptors (Hu moments, Zernike moments and Fourier descriptors), reporting the accuracy and the true positive rate for each experiment. We used Fleiss' Kappa Coefficient to evaluate the inter-expert agreement and Fisher's exact test for inter-expert variability and statistical significant differences between descriptors and learning techniques. RESULTS: Our results confirm the high degree of inter-expert variability in the morphological sperm analysis. Regarding the classification base line, we show that none of the standard descriptors or classification approaches is best suitable for tackling the problem of sperm head classification. We discovered that the correct classification rate was highly variable when trying to discriminate among non-normal sperm heads. By using the Fourier descriptor and SVM, we achieved the best mean correct classification: only 49%. CONCLUSIONS: We conclude that the SCIAN-MorphoSpermGS will provide a standard tool for evaluation of characterization and classification approaches for human sperm heads. Indeed, there is a clear need for a specific shape-based descriptor for human sperm heads and a specific classification approach to tackle the problem of high variability within subcategories of abnormal sperm cells.

Gold-standard and improved framework for sperm head segmentation.

Semen analysis is the first step in the evaluation of an infertile couple. Within this process, an accurate and objective morphological analysis becomes more critical as it is based on the correct detection and segmentation of human sperm components. In this paper, we present an improved two-stage framework for detection and segmentation of human sperm head characteristics (including acrosome and nucleus) that uses three different color spaces. The first stage detects regions of interest that define sperm heads, using k-means, then candidate heads are refined using mathematical morphology. In the second stage, we work on each region of interest to segment accurately the sperm head as well as nucleus and acrosome, using clustering and histogram statistical analysis techniques. Our proposal is also characterized by being fully automatic, where a user intervention is not required. Our experimental evaluation shows that our proposed method outperforms the state-of-the-art. This is supported by the results of different evaluation metrics. In addition, we propose a gold-standard built with the cooperation of a referent expert in the field, aiming to compare methods for detecting and segmenting sperm cells. Our results achieve notable improvement getting above 98% in the sperm head detection process at the expense of having significantly fewer false positives obtained by the state-of-the-art method. Our results also show an accurate head, acrosome and nucleus segmentation achieving over 80% overlapping against hand-segmented gold-standard. Our method achieves higher Dice coefficient, lower Hausdorff distance and less dispersion with respect to the results achieved by the state-of-the-art method.