Deep Learning for Human Embryo Classification at the Cleavage Stage (Day 3)

Zeman, Astrid; Maerten, Anne-Sofie; Mengels, A; Sharon, Lie Fong; Spiessens, Carl; Beeck, Hans Op de

doi:10.1007/978-3-030-68763-2_21

Cited by 5 publications

(5 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Zeman et al 11 proposed a branched network to predict three relevant scores of fragmentation, asymmetry, and the number of blastomeres, and then used the obtained scores to predict embryo destiny (discard vs. transfer) by a hidden layer. They only reported that the test performance reached 83.64%, but did not mention the meaning of 83.64%.…”

Section: Comparison With Existing Methods Of Embryo Classificationmentioning

confidence: 99%

“…Convolutional neural network (CNN) is a typical method of automated extracting features by use of 2D or 3D convolution in a learning step, and it has achieved great success in computer vision and image processing. [6][7][8] Inspired by the remarkable successes of CNNs, several CNN-based systems [9][10][11][12][13][14][15] have been proposed for classifying and assessing human embryos. Wu et al 9 in our group first employed the DenseNet169, Inception V3, ResNet50, and VGG19 to classify the embryos.…”

Section: Introductionmentioning

confidence: 99%

“…Thirumalaraju et al 10 used the multiple CNN architectures (Inception-v3, ResNet-50, Inception-ResNet-v2, NASNetLarge, ResNetXt-101, ResNeXt-50, and Xception) to classify the embryos, and the ResNeXt-101 achieved the highest performance with the accuracy of 90.75%. Zeman et al 11 used the standard Alexnet architecture to predict the destiny of embryos (discard or transfer), and they also used a regression model based on the extracted relevant scores to predict the embryo classification. Khosravi et al 12 developed a framework (STORK) based on the Inception model to predict blastocyst as good or poor quality, and they achieved very high performance with an accuracy of 96.94%.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Cascaded networks for the embryo classification on microscopic images using the residual external‐attention

Guo

Liu

Gong

et al. 2022

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Embryo assessment and selection are usually based on the visual morphological analysis by expert embryologists. Although the embryologist assessment has been routinely used in clinical practice, it is highly dependent on the embryologist's experience and is very time‐consuming. Therefore, objective and efficient methods for automated embryo evaluation are in high demand. We proposed a framework of cascaded networks to hierarchically extract and integrate the microscopic image features for embryo classification. The cascaded networks consisted of a coarse network and a refined network. The coarse network produced a classification activation mapping (CAM) with the highest classification probability, which indicated the most discriminative regions of embryo classification. The refined network extracted and integrated the image features again by using both the CAMs and the corresponding original images. In addition, the residual external‐attention block (ResEA) was used in the refined network to better capture long‐range dependencies. Our cascaded networks were trained on a dataset of 7728 microscopic images of day 3 embryos from 1800 couples and evaluated on an independent testing dataset of 734 microscopic images. The accuracy, sensitivity, specificity, precision, and F1‐score were employed to evaluate the performance of our cascaded networks. Compared with the coarse network and the refined network, respectively, the cascaded networks without the ResEA improved the classification results of embryos. The ResEA block helped the cascaded networks to further improve all five metrics for better embryo classification. Our proposed cascaded networks also achieved better classification results than a junior embryologist did. The cascaded networks hierarchically make full use of image features for more effective learning, and the ResEA further improves the performance of embryo classification.

show abstract

Section: Comparison With Existing Methods Of Embryo Classificationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Cascaded networks for the embryo classification on microscopic images using the residual external‐attention

Guo

Liu

Gong

et al. 2022

Int J Imaging Syst Tech

View full text Add to dashboard Cite

show abstract

“…The detailed experimental parameters are listed in Table 1. The AlexNet architecture has been proved to be the best for day 3 embryo classification [9]. In this paper, we also used this architecture on our datasets of day 3 embryo for comparison.…”

Section: 2neural Network Models Used In Our Experimentsmentioning

confidence: 99%

“…Compared with other networks, Xception performed best in embryo classification with an accuracy of 64.95% in 5 classes and 90.97% in 2 classes. Astrid Zeman et al [9] used AlexNet [10] to classify cleavage-stage embryos into transfer and discard, and achieved an accuracy of 75.24%, which was superior to the state-of-art embryo classification model, STORK, at that time on their datasets. This indicates that AlexNet is a better choice for the task of classifying day 3 embryos.…”

Section: Introductionmentioning

confidence: 99%