Automatic Assessment of Human Sperm Images with Capsule Networks

Yıl 2021, Cilt: 3 Sayı: Special Issue: Full Papers of 2nd International Congress of Updates in Biomedical Engineering, 62 - 70, 13.01.2021

Mert Şen Hatice Doğan

Öz

Infertility which is a psychologically threatening and emotionally stressful problem is seen approximately 15% of couples in worldwide. Recent studies have shown that in 40-50% of couples evaluated for infertility, the problem is caused by the male individual. Sperm morphology analysis that provides separation of normal and abnormal sperm is very important in evaluating male infertility and showing the causes. Since manual evaluation of sperm morphology is time consuming and subjective, automatic assessment methods are needed. In this study, Capsule Networks, a special model of Deep Neural Networks (DNN), are used for the classification of human sperm head images. The classification performances of capsule networks are measured using the Modified Human Sperm Morphology Analysis dataset (MHSMA). The results show that the best classification accuracy is achieved as 73%.

Anahtar Kelimeler

Capsule networks, infertility, sperm morphology

Kaynakça

• 1. Abbirany, V., V. Shanti. “Spermatoza segmentation and morphological parameter analysis based detection of teratozoospermia”. Int. J. Comput. Appl. 3(2010): 19-23.
• 2. Alegre, E., M. Biehl, N. Petkov, L. Sanchez. “Automatic classification of the acrosome status of boar spermatozoa using digital image processing and LVQ”. Comput. Biol. Med. 38(2008): 461- 468.
• 3. Bijar, A., A. P Benavent, M. Mikaeili, R. Khayati. “Fully automatic identification and discrimination of sperm’s parts in microscopic images of stained human semen smear”. J. Biomed. Sci. Eng. 5(2012): 384.
• 4. Chang, V., M. Saavedra, V. Casta𝑛̈eda, L. Sarabia, N. Hitschfeld, S. H𝑎̈rtel. “Gold-standard and improved framework for sperm head segmentation”. Comput. Methods Progr. Biomed. 117(2014): 225-237.
• 5. Google. “CapsNet-Keras”. Last update 5 August 2009. https:// github.com/XifengGuo/ CapsNet- Keras
• 6. Javadi, S., S. A. Mirroshandel. “A novel deep learning method for automatic assessment of human sperm images”. Computers in Biology and Medicine 109(2019): 182-194.
• 7. Li, J., K. K. Tseng, H. Dong, Y. Li, M. Zhao, M. Ding. “Human sperm health diagnosis with principal component analysis and k-nearest neighbor algorithm”. Medical Biometrics, 2014 International Conference on IEEE, (2014): 108-113.
• 8. R. Menkveld, C. A. Holleboom, , J. P. Rhemrev. “Measurement and significance of sperm morphology”. Asian J. Androl. 13(2011): 59.
• 9. Google. “Understanding Hinton’s capsule networks. part I: intuition”. Last update 22 December 2019. https://medium.com/ai%C2%B3-theory-practice-business/under standing-hintons-capsule- networks-part-i-intuition-b4b559d1159b
• 10.Google. “Understanding capsule network architecture”. Last update 15 December 2019. https://software.intel.com/en-us/articles/understanding-capsule-network-architecture
• 11. Sabour, S., N. Frosst, G. Hinton. “Dynamic routing between capsules”. 31st Conference on Neural Information Processing Systems (NIPS), (2017): 1-11.