EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysis
Background: Evaluating embryo quality is crucial for the success of in vitro fertilization procedures. Traditional methods, such as the Gardner grading system, rely on subjective human assessment of morphological features, leading to potential inconsistencies and errors. Artificial intelligence-powe...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
The Korean Society of Animal Reproduction and Biotechnology
2024-12-01
|
| Series: | Journal of Animal Reproduction and Biotechnology |
| Subjects: | |
| Online Access: | https://www.e-jarb.org/journal/view.html?uid=2713&vmd=Full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841536587335204864 |
|---|---|
| author | Vincent Jaehyun Shim Hosup Shim Sangho Roh |
| author_facet | Vincent Jaehyun Shim Hosup Shim Sangho Roh |
| author_sort | Vincent Jaehyun Shim |
| collection | DOAJ |
| description | Background: Evaluating embryo quality is crucial for the success of in vitro fertilization procedures. Traditional methods, such as the Gardner grading system, rely on subjective human assessment of morphological features, leading to potential inconsistencies and errors. Artificial intelligence-powered grading systems offer a more objective and consistent approach by reducing human biases and enhancing accuracy and reliability.
Methods: We evaluated the performance of five convolutional neural network architectures—EfficientNet-B0, InceptionV3, ResNet18, ResNet50, and VGG16— in grading blastocysts into five quality classes using only embryo images, without incorporating clinical or patient data. Transfer learning was applied to adapt pretrained models to our dataset, and data augmentation techniques were employed to improve model generalizability and address class imbalance.
Results: EfficientNet-B0 outperformed the other architectures, achieving the highest accuracy, area under the receiver operating characteristic curve, and F1-score across all evaluation metrics. Gradient-weighted Class Activation Mapping was used to interpret the models’ decision-making processes, revealing that the most successful models predominantly focused on the inner cell mass, a critical determinant of embryo quality.
Conclusions: Convolutional neural networks, particularly EfficientNet-B0, can significantly enhance the reliability and consistency of embryo grading in in vitro fertilization procedures by providing objective assessments based solely on embryo images. This approach offers a promising alternative to traditional subjective morphological evaluations. |
| format | Article |
| id | doaj-art-b3a1fee9f41e43a2ba1c9eb8e332b999 |
| institution | Kabale University |
| issn | 2671-4639 2671-4663 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | The Korean Society of Animal Reproduction and Biotechnology |
| record_format | Article |
| series | Journal of Animal Reproduction and Biotechnology |
| spelling | doaj-art-b3a1fee9f41e43a2ba1c9eb8e332b9992025-01-14T15:17:07ZengThe Korean Society of Animal Reproduction and BiotechnologyJournal of Animal Reproduction and Biotechnology2671-46392671-46632024-12-0139426727710.12750/JARB.39.4.267EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysisVincent Jaehyun Shim0https://orcid.org/0009-0003-2271-4776Hosup Shim1https://orcid.org/0000-0003-0329-6276Sangho Roh2https://orcid.org/0000-0001-8082-6459Cellular Reprogramming and Embryo Biotechnology Laboratory, Dental Research Institute, Seoul National University School of Dentistry, Seoul 08826, KoreaDepartment of Nanobiomedical Science, Dankook University, Cheonan 31116, KoreaCellular Reprogramming and Embryo Biotechnology Laboratory, Dental Research Institute, Seoul National University School of Dentistry, Seoul 08826, KoreaBackground: Evaluating embryo quality is crucial for the success of in vitro fertilization procedures. Traditional methods, such as the Gardner grading system, rely on subjective human assessment of morphological features, leading to potential inconsistencies and errors. Artificial intelligence-powered grading systems offer a more objective and consistent approach by reducing human biases and enhancing accuracy and reliability. Methods: We evaluated the performance of five convolutional neural network architectures—EfficientNet-B0, InceptionV3, ResNet18, ResNet50, and VGG16— in grading blastocysts into five quality classes using only embryo images, without incorporating clinical or patient data. Transfer learning was applied to adapt pretrained models to our dataset, and data augmentation techniques were employed to improve model generalizability and address class imbalance. Results: EfficientNet-B0 outperformed the other architectures, achieving the highest accuracy, area under the receiver operating characteristic curve, and F1-score across all evaluation metrics. Gradient-weighted Class Activation Mapping was used to interpret the models’ decision-making processes, revealing that the most successful models predominantly focused on the inner cell mass, a critical determinant of embryo quality. Conclusions: Convolutional neural networks, particularly EfficientNet-B0, can significantly enhance the reliability and consistency of embryo grading in in vitro fertilization procedures by providing objective assessments based solely on embryo images. This approach offers a promising alternative to traditional subjective morphological evaluations.https://www.e-jarb.org/journal/view.html?uid=2713&vmd=Fullblastocystconvolutional neural networksdeep learningembryoin vitro fertilization |
| spellingShingle | Vincent Jaehyun Shim Hosup Shim Sangho Roh EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysis Journal of Animal Reproduction and Biotechnology blastocyst convolutional neural networks deep learning embryo in vitro fertilization |
| title | EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysis |
| title_full | EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysis |
| title_fullStr | EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysis |
| title_full_unstemmed | EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysis |
| title_short | EfficientNet-B0 outperforms other CNNs in image-based five-class embryo grading: a comparative analysis |
| title_sort | efficientnet b0 outperforms other cnns in image based five class embryo grading a comparative analysis |
| topic | blastocyst convolutional neural networks deep learning embryo in vitro fertilization |
| url | https://www.e-jarb.org/journal/view.html?uid=2713&vmd=Full |
| work_keys_str_mv | AT vincentjaehyunshim efficientnetb0outperformsothercnnsinimagebasedfiveclassembryogradingacomparativeanalysis AT hosupshim efficientnetb0outperformsothercnnsinimagebasedfiveclassembryogradingacomparativeanalysis AT sanghoroh efficientnetb0outperformsothercnnsinimagebasedfiveclassembryogradingacomparativeanalysis |