Automated Sperm Assessment Framework and Neural Network Specialized for Sperm Video Recognition

TL;DR

This paper introduces a new sperm assessment framework and a specialized neural network, RoSTFine, that analyze sperm videos including morphology and motility, improving accuracy over existing models.

Contribution

The study develops a comprehensive sperm video dataset with soft-label annotations and proposes RoSTFine, a neural network tailored for sperm recognition, addressing limitations of previous image-based methods.

Findings

RoSTFine outperforms existing video recognition models in sperm assessment accuracy.

The model effectively focuses on critical sperm parts like head and neck.

Soft-label annotations improve assessment consistency.

Abstract

Infertility is a global health problem, and an increasing number of couples are seeking medical assistance to achieve reproduction, at least half of which are caused by men. The success rate of assisted reproductive technologies depends on sperm assessment, in which experts determine whether sperm can be used for reproduction based on morphology and motility of sperm. Previous sperm assessment studies with deep learning have used datasets comprising images that include only sperm heads, which cannot consider motility and other morphologies of sperm. Furthermore, the labels of the dataset are one-hot, which provides insufficient support for experts, because assessment results are inconsistent between experts, and they have no absolute answer. Therefore, we constructed the video dataset for sperm assessment whose videos include sperm head as well as neck and tail, and its labels were annotated with soft-label. Furthermore, we proposed the sperm assessment framework and the neural network, RoSTFine, for sperm video recognition. Experimental results showed that RoSTFine could improve the sperm assessment performances compared to existing video recognition models and focus strongly on important sperm parts (i.e., head and neck).