TY  - JOUR
T1  - Artificial intelligence in assisted reproductive techniques: comparative evaluation of deep learning architectures for bovine cumulus-oocyte complexes classification
TT  - Yapay zekâ destekli üreme tekniklerinde: sığır kumulus-oosit komplekslerinin sınıflandırılmasına yönelik derin öğrenme mimarilerinin karşılaştırmalı değerlendirmesi
AU  - Çavuşoğlu, Türker
AU  - Gökhan, Aylin
AU  - Çetinkaya Karabekir, Seda
AU  - Ölmez, Emre
AU  - Akarca Dizakar, Saadet Özen
AU  - Tekel, Mert Can
AU  - Er, Orhan
AU  - Güllü, Mehmet Kemal
PY  - 2025
DA  - December
Y2  - 2025
DO  - 10.19161/etd.1737365
JF  - Ege Tıp Dergisi
JO  - EJM
PB  - Ege University
WT  - DergiPark
SN  - 1016-9113
SP  - 653
EP  - 664
VL  - 64
IS  - 4
LA  - en
AB  - Aim: Follicular quality is a key determinant of success in assisted reproductive technologies, directly affecting outcomes such as fertilization, embryo development, implantation, and live birth rates. However, conventional assessment of cumulus-oocyte complexes relies on subjective morphological evaluation, introducing variability and reducing consistency in clinical decision-making.Materials and Methods: A comparative evaluation of various pre-trained deep learning architectures—including both convolutional neural networks and transformer-based models—was conducted for the automated morphological grading of bovine cumulus-oocyte complexes into four quality categories (Grade A–D). A dataset of 1,400 annotated images of cumulus-oocyte complexes, enhanced through data augmentation techniques to increase image diversity, was used for model training and validation.Results: Among the tested architectures, Xception41 (a variant of convolutional neural networks) and Swin Transformer (a transformer-based model) achieved the highest performance, with test accuracies of 74.75% and 73.25%, and macro F1-scores of 0.75 and 0.74, respectively. While both models performed well in grading cumulus-oocyte complexes with distinct morphological features (Grades 3 and 4), classification accuracy decreased for the more subtle differences between Grades 1 and 2. Furthermore, most models exhibited signs of overfitting under the current training configuration.Conclusion: This study demonstrates the potential of deep learning-based approaches to standardize and enhance the efficiency of cumulus-oocyte complexes evaluation in assisted reproductive technologies. Further optimization is needed to improve model generalization and to address challenges in grading morphologically similar follicular structures.
KW  - Artificial Intelligence
KW  - Convolutional Neural Networks
KW  - Xception41
KW  - Deep Learning
KW  - Swin Transformer
KW  - Cumulus-Oocyte Complexes
KW  - Oocyte Classification
KW  - Assisted Reproductive Technologies
N2  - Amaç: Folikül kalitesi, yardımcı üreme teknolojilerindeki başarıyı belirleyen temel bir faktördür ve döllenme, embriyo gelişimi, implantasyon ve canlı doğum oranları gibi önemli sonuçları doğrudan etkiler. Ancak, kumulus-oosit komplekslerinin geleneksel değerlendirmesi, öznel morfolojik gözlemlere dayandığı için klinik karar süreçlerinde değişkenliğe ve tutarsızlığa neden olmaktadır.Gereç ve Yöntem: Sığır kumulus-oosit komplekslerinin otomatik olarak morfolojik kaliteye göre dört kategoriye (A–D) ayrılması amacıyla, önceden eğitilmiş çeşitli derin öğrenme mimarileri—evrişimsel sinir ağları ve dönüştürücü tabanlı modeller dahil—karşılaştırmalı olarak değerlendirilmiştir. Görsel çeşitliliği artırmak amacıyla veri artırma teknikleri uygulanarak oluşturulan 1.400 etiketlenmiş kumulus-oosit kompleksi görüntülerinden oluşan veri seti, modellerin eğitimi ve doğrulaması için kullanılmıştır.Bulgular: Test edilen mimariler arasında Xception41 (evrişimsel sinir ağları varyantı) ve Swin Transformer (dönüştürücü tabanlı bir model), sırasıyla %74,75 ve %73,25 test doğrulukları ile 0,75 ve 0,74 makro F1 skorlarına ulaşarak en yüksek performansı göstermiştir. Bu modeller, belirgin morfolojik özelliklere sahip 3. ve 4. derece kumulus-oosit komplekslerinde yüksek sınıflandırma başarısı gösterirken, 1. ve 2. dereceler arasındaki daha ince farkların ayırt edilmesinde zorlanmıştır. Ayrıca, mevcut eğitim konfigürasyonu altında çoğu modelde aşırı öğrenme eğilimi gözlemlenmiştir.Sonuç: Bu çalışma, derin öğrenme tabanlı yaklaşımların yardımcı üreme teknolojileri kapsamında kumulus-oosit kompleksi değerlendirmesini standartlaştırma ve değerlendirme süreçlerinin etkinliğini artırma potansiyelini ortaya koymaktadır. Bununla birlikte, morfolojik olarak benzer foliküler yapılar arasındaki sınıflandırma zorluklarının aşılması ve model genelleme yeteneğinin artırılması için ilave optimizasyon gereklidir.
CR  - Fainberg J, Kashanian JA. Recent advances in understanding and managing male infertility. F1000Res. 2019;8:670. doi:10.12688/f1000research.17076.1
CR  - Crawford S, Fussman C, Bailey M, Bernson D, Jamieson DJ, Murray-Jordan M, et al. Estimates of lifetime infertility from three states: The Behavioral Risk Factor Surveillance System. J Womens Health (Larchmt). 2015;24(7):578-86. doi:10.1089/jwh.2014.5102
CR  - Čegar B, Šipetić Grujičić S, Bjekić J, Vuksanović A, Bojanić N, Bartolović D, et al. Understanding the male perspective: evaluating quality of life and psychological distress in Serbian men undergoing infertility treatment. Life (Basel). 2023;13(9):1894. doi:10.3390/life13091894
CR  - Uhde K, van Tol HTA, Stout TAE, Roelen BAJ. Metabolomic profiles of bovine cumulus cells and cumulus-oocyte-complex-conditioned medium during maturation in vitro. Sci Rep. 2018;8:9477. doi:10.1038/s41598-8-27829-9
CR  - Richani D, Dunning KR, Thompson JG, Gilchrist RB. Metabolic co-dependence of the oocyte and cumulus cells: essential role in determining oocyte developmental competence. Hum Reprod Update. 2021;27(1):27-47. doi:10.1093/humupd/dmaa025
CR  - Choi Y, Moon SH. Types and characteristics of stress coping in women undergoing infertility treatment in Korea. Int J Environ Res Public Health. 2023;20(3):2648. doi:10.3390/ijerph20032648
CR  - Stojkovic M, Machado SA, Stojkovic P, Zakhartchenko V, Hutzler P, Gonçalves PB, Wolf E Mitochondrial distribution and adenosine triphosphate content of bovine oocytes before and after in vitro maturation: correlation with morphological criteria and developmental capacity after in vitro fertilization and culture. Biol Reprod. 2001;64(3):904-909
CR  - Hartmann W, Pereira JFS. Biotechnics applied to bovine female. In: Bergstein-Galan TG, editor. Reproduction Biotechnology in Farm Animals. Brazil: AvidScience; 2018. p. 155-180
CR  - Cavusoglu T, Gokhan A, Sirin C, Tomruk C, Kilic KD, Olmez E, et al. Classification of bovine cumulus-oocyte complexes with convolutional neural networks. Med Records. 2023;5(3):489-5. doi:10.37990/medr.1292782
CR  - Chollet F. Xception: deep learning with depthwise separable convolutions. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway (NJ): IEEE; 2017. p. 1800-7. doi:10.1109/CVPR.2017.195
CR  - Chen LC, Zhu Y, Papandreou G, Schroff F, Adam H. Encoder-decoder with atrous separable convolution for semantic image segmentation [Internet]. arXiv; 2018 Feb 7 [cited 2025 Aug 10]. Available from: http://arxiv.org/abs/1802.02611.
CR  - Liu Z, Mao H, Wu CY, Feichtenhofer C, Darrell T, Xie S. A ConvNet for the 2020s [Internet]. arXiv; 2022 Jan 10 [cited 2025 Aug 10]. Available from: http://arxiv.org/abs/2201.03545.
CR  - Chen LC, Zhu Y, Papandreou G, Schroff F, Adam H. Encoder-decoder with atrous separable convolution for semantic image segmentation. In: Ferrari V, Hebert M, Sminchisescu C, Weiss Y, editors. Computer Vision – ECCV 2018. Lecture Notes in Computer Science, vol. 11211. Cham: Springer; 2018. p. 833-51. doi:10.1007/978-3-030-01234-2_49
CR  - Park W, Schwendicke F, Krois J, Huh JK, Lee JH. Identification of dental implant systems using a large-scale multicenter data set. J Dent Res. 2023;102(7):727-33. doi:10.1177/00220345231160750
CR  - Huang G, Liu Z, van der Maaten L, Weinberger KQ. Densely connected convolutional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway (NJ): IEEE; 2017. p. 2261-9. doi:10.1109/CVPR.2017.243
CR  - Liao Q, Zhang Q, Feng X, Huang H, Xu H, Tian B, et al. Development of deep learning algorithms for predicting blastocyst formation and quality by time-lapse monitoring. Commun Biol. 2021;4:415. doi:10.1038/s42003-021-01937-1
CR  - Dosovitskiy A, Beyer L, Kolesnikov A, Weissenborn D, Zhai X, Unterthiner T, et al. An image is worth 16x16 words: transformers for image recognition at scale [Internet]. arXiv; 2020 Oct 23 [cited 2025 Aug 10]. Available from: http://arxiv.org/abs/2010.11929
CR  - Savaş S. Application of deep ensemble learning for palm disease detection in smart agriculture. Heliyon. 2024;10(17):e37141. doi:10.1016/j.heliyon.2024.e37141
CR  - Bao H, Dong L, Piao S, Wei F. BEiT: BERT pre-training of image transformers [Internet]. arXiv; 2021 Jun 15 [cited 2025 Aug 10]. Available from: http://arxiv.org/abs/2106.08254
CR  - ElMoaqet H, Janini R, Ryalat M, Al-Refai G, Abdulbaki Alshirbaji T, Jalal NA, et al. Using masked image modelling transformer architecture for laparoscopic surgical tool classification and localization. Sensors (Basel). 2025;25(10):3017. doi:10.3390/s25103017
CR  - Kong F, Shi Z, Cao H, Hao Y, Cao Q. Self-supervised U-transformer network with mask reconstruction for metal artifact reduction. Phys Med Biol. 2025;70(6):065009. doi:10.1088/1361-6560/adbaae
CR  - Liu Z, Lin Y, Cao Y, Hu H, Wei Y, Zhang Z, et al. Swin transformer: hierarchical vision transformer using shifted windows. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). Piscataway (NJ): IEEE; 2021. p. 9992-10002. doi:10.1109/ICCV48922.2021.00986
CR  - Omer AAM. Image classification based on vision transformer. J Comput Commun. 2024;12(4):49-59. doi:10.4236/jcc.2024.124005
CR  - Li Y, Riganello F, Yu J, Vatrano M, Shen M, Cheng L, et al. The autonomic response following taVNS predicts changes in level of consciousness in DoC patients. Sci Rep. 2025;15:7317. doi:10.1038/s41598-024-84029-4
CR  - Ishaq M, Raza S, Rehar H, Abadeen SeZu, Hussain D, Naqvi RA, et al. Assisting the human embryo viability assessment by deep learning for in vitro fertilization. Mathematics. 2023;11(9):2023. doi:10.3390/math11092023
CR  - Arsalan M, Haider A, Choi J, Park KR. Detecting blastocyst components by artificial intelligence for human embryological analysis to improve success rate of in vitro fertilization. J Pers Med. 2022;12(2):124. doi:10.3390/jpm12020124
CR  - Mushtaq A, Mumtaz M, Raza A, Salem N, Yasir MN. Artificial intelligence-based detection of human embryo components for assisted reproduction by in vitro fertilization. Sensors (Basel). 2022;22(19):7418. doi:10.3390/s22197418
CR  - Jayashree P, Mitra S. Facilitating a deep approach to learning: an innovative case assessment technique. J Manag Organ. 2012;18(4):555-572. doi:10.5172/jmo.2012.18.4.555
CR  - Allahbadia GN. Embryo transfer is the last frontier for deep machine learning and artificial intelligence in medically assisted reproduction (MAR). J Reprod. 2023;2(1):28-38. doi:10.58779/issn.2954-467X.tjor2023.v2.n1.18.
CR  - Athanasiou G, Cerquides J, Raes A, Azari-Dolatabad N, Angel-Velez D, Van Soom A, et al. Detecting the area of bovine cumulus oocyte complexes using deep learning and semantic segmentation. In: Frontiers in Artificial Intelligence and Applications. Vol. 356, Artificial Intelligence Research and Development. Amsterdam: IOS Press; 2022. p. 249-258. doi:10.3233/FAIA220346
UR  - https://doi.org/10.19161/etd.1737365
L1  - https://dergipark.org.tr/en/download/article-file/5034130
ER  -