Learning sperm cells part segmentation with class-specific data augmentation

Infertility affects around 15% of couples worldwide. Male fertility problems include poor sperm quality and low sperm count. The advanced fertility treatment methods like ICSI are nowadays supported by vision systems to assist embryologists in selecting good quality sperm. Computer-Assisted Semen Analysis (CASA) provides quantitative and qualitative sperm analysis concerning concentration, motility, morphology, vitality, and fragmentation. However, fertility assessment algorithms often neglect individual spermatozoon tail and its beating patterns because recognizing the tails in blurry microscopic images reliably is challenging. In this article, we propose that models trained with head and tail part classes can better localize parts and segment the whole spermatozoon objects. Usually, the training of segmentation sperm models is supported by image-level augmentation. We argue that models guided by class-specific data augmentation attend to less discriminative sperm parts. To demonstrate this, we decouple the augmentation into object-level and background augmentation for the sperm part segmentation problem. Our proposed method outperforms state-of-the-art methods on the SegSperm dataset. Moreover, our ablation studies confirm the effectiveness of the proposed part-based object representation and augmentation.

I. Introduction

The problem of full sperm segmentation is challenging for various reasons [1], [2]. The depth of field of optical microscopes is insufficient to keep entire spermatozoons in focus, especially the quickly moving, thin and long flagella. A typical tail is 45 µm long with sharp bend and coils [3], [4]. Sperm tail vary: they can be short, multiple, broken, bent, irregular in shape, coiled, or with any combination of these attributes [4]. Sperm heads are usually less blurry. While the motion of spermatozoa head can be graded as progressively motile, the tail motion is a superposition of progressive motile and flagellar beating pattern. Current computer vision tools for sperm anal ysis do not provide details related to flagellar beating forces, which have been little studied [5]. Additional challenges for segmentation include background noise and artifacts resembling sperm head. The segmentation in the presence of noise is a common problem for biomedical imaging [6] and especially microscopy imaging [7]. The final challenge is the time efficiency of the segmentation method. To assist the embryologist in selecting the better spermatozoon for injection in an ICSI procedure, the method needs to run in real-time [8], [9] at video rates.