A spectral framework for sperm shape characterization

doi:10.1016/j.compbiomed.2004.03.007

Computers in Biology and Medicine

Volume 35, Issue 6, July 2005, Pages 463-473

https://doi.org/10.1016/j.compbiomed.2004.03.007 Get rights and content

Abstract

A novel methodology for characterization of animal sperm shape involving the use of a spectral approach to multiscale curvature estimation is proposed. By using the derivative property of the Fourier transform, allied to Gaussian smoothing, accurate estimates of the curvature along the sperm contour can be obtained in such a way that the curvature peaks corresponding to the sperm head vertices can be effectively identified. The measurements derived from such a processing, namely the width of the basal region of the head, the centralization of tail implantation, and the multiscale bending energy, provide valuable resources for fertility and phylogenetic studies.

Introduction

Aimed at optimizing the speed and accuracy while avoiding the intrinsic subjectivity of human operators, computational semen analysis has been increasingly considered as a routine procedure in several research and diagnostic laboratories. Originally oriented to sperm motility, several software applications can be found currently that are capable of performing additional morphological analysis in automated or semi-automated fashion. While many such programs concentrate on the geometric properties of the sperm head, some of them also allow the characterization of alterations in the sperm tail, identifying total length, irregular shapes (bent and coiled tail), as well as the presence of cytoplasm droplets [1], [2], [3].

Several measurements have been considered for sperm morphology characterization and analysis. Typically, off-the-shelf programs estimate traditional basic features such as area, perimeter, length and width, as well as derived measurements including width–length ratio, roundness, ellipticity and shape factor, among others. More sophisticated geometrical features such as Fourier descriptors and symmetry have also been considered [4].

As human sperm heads are predominantly elliptical, little attention is generally given to the placement of tail implantation. Because several of the currently available software programs for animal sperm analysis are adaptations of human sperm analysis software, few resources are usually available for the characterization of tail implantation. At the same time, several commercial species such as bovine, rams and ovine present an oval head with a plan extremity (head basis) with characteristic central tail implantation. Displaced implantations (abaxial) generally interfere with sperm motility, with obvious implications for fertility [5]. Another typical abnormality is the narrowing of the head basis [5]. As a consequence of difficulties for properly identifying the basal region and the tail implantation position, the estimation of geometrical measurements characterizing such important alterations is seldom available in commercial software.

Morphometric characterization of the sperm head is frequently used for male fertility evaluation in human and veterinary medicine, being frequently enhanced by application of computational means. Detailed study of these features suggests the existence of correlations between the phylogeny and sperm shape [6]. Combinations of the morphometric features make possible separation of spermatozoa of different species. Such combinations generally possess high sensitivity in the identification of the carrying spermatozoa morphologic abnormalities, therefore representing an important resource for the evaluation of male fertility.

Introduced by Costa [7], [8], the spectral approach to multiscale curvature estimation provides a natural means for the identification and characterization of sperm morphology. The method is based on the derivative property of the Fourier transform, which is applied over smoothed (i.e. low-pass filtered) versions of the contour of the object. As a Gaussian function is used as low-pass filter kernel, its standard deviation defines a spatial scale parameter that can be used to control the level of detail considered in the analysis, allowing effective removal of small scale noise implied by the spatial quantization of the object when captured as a digital image. The special importance of the curvature for vertex identification stems from the fact that these features of the object contour are characterized by peaks of curvature [9]. As such, the technique allows a natural means to identify not only the placement of the tail implantation, but also the two vertices characterizing the sperm head basal region. In addition, the multiscale bending energy [10] of the frontal portion of the sperm head contour can be obtained as an immediate byproduct of the methodology. The bending energy provides an interesting conceptual interpretation as being proportional to the amount of energy one has to apply to a circle in order to transform it into the object under analysis.

The present article describes how the multiscale curvature and the closely related multiscale bending energy, numerically estimated by a spectral method, can be used for sperm shape analysis, more specifically as a means of quantifying the tail implantation, width of basal region, and energy of the frontal portion of the sperm head.

This paper starts with a description of the multiscale curvature and its application to sperm morphometric characterization (Section 2.1), making possible the vertex head detection (Section 2.2). Then we present new morphometric measures (Section 2.3) that, when applied to bovine, caprine, ovine and canine sperm are evaluated considering diverse combinations. These measurements allowed the descrimination of several species of spermatozoa, as well as respective subpopulations.

Section snippets

Spectral curvature estimation

The sperm contour is henceforth assumed to be represented as the parametric curve defined by the Cartesian coordinates (x(s),y(s)) of each of its elements, where s is the arc-length parameter. Such a representation can be obtained by using standard cell segmentation and contour following algorithms such as those described in [9], [11]. The curvature of a parametric contour is formally defined as in Eq. (1), which is irrespective of the adopted parameterization $k(s)= x ̇ (s) y ̈ (s)− y ̇ (s) x ̈ (s) (x ̇ (s)^{2} + y ̇$

Experimental results

The methods used for identification of the sperm head vertices are verified to be generally effective, as demonstrated in Fig. 1. The results were plotted considering two-by-two combinations of the measurements, as shown in Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8. Fig. 3 shows the cluster distribution obtained considering ε(σ=3) and ε(σ=5), allowing good separation between the species, mainly along the y-axis (related to ε(σ=5)), as well as the presence of spermatozoa displaced from the

Conclusions

This article reports on the application of a spectral approach for multiscale curvature estimation to the characterization of animal sperm head geometry, with special attention given to the width of the basal portion, centralization of tail implantation, and energy of the frontal portion of the head. Given their potential for diagnosis and phylogenetic investigations, such measures can assist investigations on animal fertility, which had been previously constrained by most of the existing

Summary

The characterization of sperm morphology is of great importance for objective and quantitative analysis of fertility of individuals, as well as for phylogenetic studies. The present work describes how a spectral approach to derivative estimation has been used to obtain particularly relevant measurements of sperm head morphology. The methodology involves extracting the sperm head outlines and using the Fourier transform to estimate its first and second derivatives, from which the respective

Acknowledgements

Luciano da F. Costa and Marcelo E. Beletti are grateful to Conselho Nacional de Desenvolvimento Cientı́fico e Tecnológico - CNPq (308231/03-1, 150020/02-3 and 350058/03-2) and Fundação de Amparo à Pesquisa do Estado de São Paulo - FAPESP (99/12765-2) for financial support.

Marcelo Emı́lio Beletti is an Adjunct Professor at the Federal University at Uberlândia, MG, PhD at the Cell Biology, and has 5 papers in international journals. His principal interests include Theriogenology, Biological Morphology and Image Analysis.

References (12)

R.M. Cesar et al.
Piecewise linear segmentation of digital contours in O(N.LOG(N)) trough technique based on effective digital curvature estimation
Real-Time Imaging
(1995)
T. Suzuki et al.
Comparison of the Sperm Quality Analyzer IIC variables with the computer-aided sperm analysis estimates
Int. J. Androl.
(2002)
J.G. Wijchman et al.
Variation in semen parameters derived from computer-aided semen analysis, within donors and between donors
J. Androl.
(2001)
G.C. Ostermeier et al.
Relationship of bull fertility to sperm nuclear shape
J. Androl.
(2001)
G.C. Ostermeier et al.
Measurements of bovine sperm nuclear shape using Fourier harmonic amplitudes
J. Androl.
(2001)
B. Hafez et al.
Reproduction in Farm Animals
(2000)

There are more references available in the full text version of this article.

Cited by (14)

Automatic classification of human sperm head morphology
2017, Computers in Biology and Medicine
Citation Excerpt :
About morphological characterization of sperm heads, Beletti et al. [33] described the morphology of sperm heads using features such as: head area, perimeter, width, length, aspect ratio, ellipticity, shape factor, width of sperm basis, the three first Fourier values, side and anterior-posterior symmetry and hydrodynamic coefficient. In [32], the authors characterized the animal sperm head shape using a multiscale curvature estimation. They showed how a spectral approach to estimate the derivative property of the Fourier transform is useful for extracting relevant features for sperm head morphology assessment.
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.
Assessment of acrosome state in boar spermatozoa heads using n-contours descriptor and RLVQ
2013, Computer Methods and Programs in Biomedicine
Citation Excerpt :
Following this idea, a number of digital image processing methods have been suggested for the segmentation, description and classification of sperm cells. Examples of this are the publications of Beletti et al., where the authors describe the spermatozoa head's morphology [9], while in [10], they characterize the animal sperm shape using a multiscale curvature estimation. Other authors, such as Linneberg et al. [11] describe the human sperm head morphology using invariant Fourier descriptors grouped in energy bands classifying them by means of neural networks.
This paper proposes a method for assessing the acrosome state of boar spermatozoa heads using digital image processing. We use gray level images in which spermatozoa have been labeled as acrosome-intact or acrosome damaged using the information of a coupled fluorescent image. The heads are segmented obtaining the outer head contour. A set of “n” inner contours separated by a logarithmic distance function is calculated later. For each point of the, in this case, seven contours a number of local texture features are computed. We have compared the classification performance of Relevance Learning Vector Quantization, class conditional means and KNN, employing cross-validation for the evaluation. Gradient magnitude data offer the best result with an overall test error of only 1%. This result outperforms previously applied methods and suggests this approach as an interesting automatized approach to this veterinarian problem.
Texture and moments-based classification of the acrosome integrity of boar spermatozoa images
2012, Computer Methods and Programs in Biomedicine
Citation Excerpt :
These approaches were first designed for human semen analysis [1], but they have been currently adapted to other species [2]. These systems are essentially based on parameters such as motility or morphometry [3–6], as they are directly related to the semen quality. However, up to our knowledge, the evaluation of the acrosome integrity of the spermatozoon heads is carried out manually, using stains and there are not any computer assisted tools for that analysis.
The automated assessment of the sperm quality is an important challenge in the veterinary field. In this paper, we explore how to describe the acrosomes of boar spermatozoa using image analysis so that they can be automatically categorized as intact or damaged. Our proposal aims at characterizing the acrosomes by means of texture features. The texture is described using first order statistics and features derived from the co-occurrence matrix of the image, both computed from the original image and from the coefficients yielded by the Discrete Wavelet Transform. Texture descriptors are evaluated and compared with moments-based descriptors in terms of the classification accuracy they provide. Experimental results with a Multilayer Perceptron and the k-Nearest Neighbours classifiers show that texture descriptors outperform moment-based descriptors, reaching an accuracy of 94.93%, which makes this approach very attractive for the veterinarian community.
Automatic classification of the acrosome status of boar spermatozoa using digital image processing and LVQ
2008, Computers in Biology and Medicine
We consider images of boar spermatozoa obtained with an optical phase-contrast microscope. Our goal is to automatically classify single sperm cells as acrosome-intact (class 1) or acrosome-damaged (class 2). Such classification is important for the estimation of the fertilization potential of a sperm sample for artificial insemination. We segment the sperm heads and compute a feature vector for each head. As a feature vector we use the gradient magnitude along the contour of the sperm head. We apply learning vector quantization (LVQ) to the feature vectors obtained for 320 heads that were labelled as intact or damaged using stains. A LVQ system with four prototypes (two for each class) allows us to classify cells with an overall test error of 6.8%. This is considered to be sufficient for semen quality control in an artificial insemination center.
Cut-off values for normal sperm morphology and toxicology for automated analysis of rat sperm morphology and morphometry
2018, Biotechnic and Histochemistry
Using sperm morphometry and multivariate analysis to differentiate species of gray Mazama
2016, Royal Society Open Science

View all citing articles on Scopus

Luciano da Fontoura Costa is an Associate Professor at the Institute of Physics at São Carlos, University of São Paulo, where he coordinate the Cybernetic Vision Research Group, which he founded in 1993. He has about 100 papers in international journals and belongs to the editorial board of Journal of Real-Time Imaging (Elsevier), EURASIP Applied Signal Processing (EURASIP-Hindawi), Journal of Intelligent and Fuzzy Systems (IOS), Psyche (The MIT Press), Neuroinformatics (Humana Press) and Journal of Integrative Neuroscience (World Scientific), being one of the Associate Editors of the latter. He has authored, jointly with Roberto M. Cesar Jr, the book Shape Analysis and Classification (CRC Press, 2001). His main interests include Neuroinformatics, Bioinformatics, Computer Vision, Signal Processing, Physical Modeling and Pattern Recognition.

Matheus Palhares Viana is an undergrad student at the Physics Department, Federal University at São Carlos, and is interested in Image Analysis and Physical Applications.

View full text

A spectral framework for sperm shape characterization

Abstract

Introduction

Section snippets

Spectral curvature estimation

Experimental results

Conclusions

Summary

Acknowledgements

Real-Time Imaging

Comparison of the Sperm Quality Analyzer IIC variables with the computer-aided sperm analysis estimates

Int. J. Androl.

Variation in semen parameters derived from computer-aided semen analysis, within donors and between donors

J. Androl.

Relationship of bull fertility to sperm nuclear shape

J. Androl.

Measurements of bovine sperm nuclear shape using Fourier harmonic amplitudes

J. Androl.

Reproduction in Farm Animals