Effective encoder-decoder network for pupil light reflex segmentation in facial photographs of ptosis patients.
Facial photograph
Ptosis
Pupil light reflex
Segmentation
Journal
Scientific reports
ISSN: 2045-2322
Titre abrégé: Sci Rep
Pays: England
ID NLM: 101563288
Informations de publication
Date de publication:
31 Oct 2024
31 Oct 2024
Historique:
received:
29
04
2024
accepted:
18
10
2024
medline:
1
11
2024
pubmed:
1
11
2024
entrez:
1
11
2024
Statut:
epublish
Résumé
Accurate segmentation of pupil light reflexes is essential for the reliable assessment of ptosis severity, a condition characterized by the drooping of the upper eyelid. This study introduces a novel encoder-decoder network specialized in reflex segmentation by focusing on addressing issues related to very small regions of interest from an architectural perspective. Specifically, the proposed network is designed to exploit low-level features effectively by integrating a multi-level skip connection and a 1 × 1 convolution-enhanced initial encoding stage. Assessed using a photograph image dataset from Chung-Ang University Hospital, which includes 87 healthy subjects, 64 with ptosis, and 257 with Graves' orbitopathy (collected between January 2010 and February 2023), the proposed network outperforms five conventional encoder-decoders. Over 30 trials, the proposed network achieved a mean Dice coefficient of 0.767 and an Intersection over Union of 0.653, indicating a statistically significant improvement in the segmentation of reflex. Our findings show that an elaborate design based on the lowest-level skip connection and 1 × 1 convolution at initial stage enhances the segmentation of pupil light reflexes. The source code of the proposed network is available at https://github.com/tkdgur658/ReflexNet .
Identifiants
pubmed: 39482369
doi: 10.1038/s41598-024-77001-9
pii: 10.1038/s41598-024-77001-9
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
26220Subventions
Organisme : Ministry of Science and ICT, South Korea
ID : 2021-0-01341
Organisme : Ministry of Science and ICT, South Korea
ID : 2021-0-01341
Organisme : Ministry of Science and ICT, South Korea
ID : 2021-0-01341
Organisme : National Research Foundation of Korea
ID : 2021R1A2C1011351
Informations de copyright
© 2024. The Author(s).
Références
Shen, J. Q., Li, H. Y., Chen, Y. H., Liu, L. & Cui, H. G. Clinical observations of corneal topographic and tomographic changes in congenital ptosis eyes: a study in China. Int. Ophthalmol. 43, 1581–1590. https://doi.org/10.1007/s10792-022-02557-2 (2023).
doi: 10.1007/s10792-022-02557-2
pubmed: 36269442
Nowak-Gospodarowicz, I., Kicinska, A., Kinasz, M. & Rekas, M. A new algorithm for the transconjunctival correction of moderate to severe upper eyelid ptosis in adults. Sci. Rep. 14,2566. https://doi.org/10.1038/s41598-024-52990-9 (2024).
Agha, M., Ismail, H., Sawaya, R. & Salameh, J. Efficacy of apraclonidine eye drops in treating ptosis secondary to myasthenia gravis: a pilot clinical trial. Muscle Nerve. 68, 206–210. https://doi.org/10.1002/mus.27851 (2023).
doi: 10.1002/mus.27851
pubmed: 37259693
Rana, K. et al. Normal periocular anthropometric measurements in an Australian population. Int. Ophthalmol. 43, 2695–2701. https://doi.org/10.1007/s10792-023-02669-3 (2023).
doi: 10.1007/s10792-023-02669-3
pubmed: 36869978
pmcid: 10371930
Alkeswani, A., Hataway, F., Westbrook, B., Gulamani, S. & Collawn, S. S. Changes in lid crease measurements in Levator Advancement for Ptosis. Ann. Plas Surg. 84, S358–S360. https://doi.org/10.1097/Sap.0000000000002304 (2020).
doi: 10.1097/Sap.0000000000002304
Song, B. et al. Introduction of deep learning-based infrared image analysis to marginal reflex distance1 measurement method to simultaneously capture images and compute results: clinical validation study. J. Clin. Med. 12, 7466. https://doi.org/10.3390/jcm12237466 (2023).
Yolcu, D. & Ozdogan, S. A novel method to measure margin reflex distance using the autorefractometer. Int. Ophthalmol. 42, 1241–1247. https://doi.org/10.1007/s10792-021-02110-7 (2022).
doi: 10.1007/s10792-021-02110-7
pubmed: 34738206
Ulku, I. & Akagündüz, E. A survey on deep learning-based architectures for semantic segmentation on 2D images. Appl. Artif. Intell. 36, 2032924. https://doi.org/10.1080/08839514.2022.2032924 (2022).
Wang, R. S. et al. Medical image segmentation using deep learning: a survey. Iet Image Process. 16, 1243–1267. https://doi.org/10.1049/ipr2.12419 (2022).
doi: 10.1049/ipr2.12419
Han, S. Y., Kwon, H., Kim, Y. & Cho, N. I. Noise-robust pupil center detection through CNN-based segmentation with shape-prior loss. Ieee Access. 8, 64739–64749. https://doi.org/10.1109/Access.2020.2985095 (2020).
doi: 10.1109/Access.2020.2985095
Yiu, Y. H. et al. DeepVOG: open-source pupil segmentation and gaze estimation in neuroscience using deep learning. J. Neurosci. Meth. 324, 108307. https://doi.org/10.1016/j.jneumeth.2019.05.016 (2019).
Song, B. et al. Novel method to measure marginal reflex distance-1 (MRD-1) using based on deep learning method. In Proceedings of the 2023 IEEE Conference on Artificial Intelligence (CAI). 316–318 (2023).
Shao, J. et al. Deep learning-based image analysis of eyelid morphology in thyroid-associated ophthalmopathy. Quant. Imag. Med. Surg. 13, 1592. https://doi.org/10.21037/qims-22-551 (2023).
Badrinarayanan, V., Kendall, A. & Cipolla, R. SegNet: a deep convolutional encoder-decoder architecture for image segmentation. Ieee T Pattern Anal. 39, 2481–2495. https://doi.org/10.1109/Tpami.2016.2644615 (2017).
Taghanaki, S. A., Abhishek, K., Cohen, J. P., Cohen-Adad, J. & Hamarneh, G. Deep semantic segmentation of natural and medical images: a review. Artif. Intell. Rev. 54, 137–178. https://doi.org/10.1007/s10462-020-09854-1 (2021).
doi: 10.1007/s10462-020-09854-1
Han, Z. M., Jian, M. W. & Wang, G. G. ConvUNeXt: an efficient convolution neural network for medical image segmentation. Knowl-Based Syst. 253, 109512. https://doi.org/10.1016/j.knosys.2022.109512 (2022).
Xu, Q., Ma, Z. C., He, N. & Duan, W. T. DCSAU-Net: a deeper and more compact split-attention U-Net for medical image segmentation. Comput. Biol. Med. 154, 106626. https://doi.org/10.1016/j.compbiomed.2023.106626 (2023).
Cao, H. et al. Swin-Unet: Unet-like pure transformer for medical image segmentation. In Proceedings of the 17th European Conference on Computer Vision (ECCV). 205–218 (2022).
Lou, A. E., Guan, S. Y. & Loew, M. CFPNet-M: a light-weight encoder-decoder based network for multimodal biomedical image real-time segmentation. Comput. Biol. Med. 154, 106579. https://doi.org/10.1016/j.compbiomed.2023.106579 (2023).
Pourafkham, B. & Khotanlou, H. ES-Net: Unet-based model for the semantic segmentation of Iris. Multimed. Tools Appl. 1-22 (2024). https://doi.org/10.1007/s11042-024-19488-y
hen, L. C. & a., Z. Yukun and Papandreou, George and Schroff, Florian and Adam, Hartwig. Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proceedings of the 15th European Conference on Computer Vision (ECCV). 1–18 (2018)
Nam, Y., Song, T., Lee, J. & Lee, J. K. Development of a neural network-based automated eyelid measurement system. Sci. Rep. 14, 1202. https://doi.org/10.1038/s41598-024-51838-6 (2024).
Giap, B. D. et al. Adaptive Tensor-based feature extraction for pupil segmentation in cataract surgery. Ieee J. Biomed. Health. 28, 1599–1610. https://doi.org/10.1109/Jbhi.2023.3345837 (2024).
doi: 10.1109/Jbhi.2023.3345837
Lou, A., Loew, M. & Cfpnet Channel-wise Feature pyramid for real-time semantic segmentation. Ieee Image Proc. 1894–1898. https://doi.org/10.1109/Icip42928.2021.9506485 (2021).
Loshchilov, I. & Hutter, F. Decoupled weight decay regularization. in Proc. Int. Conf. Learn. Represent. (2019).
Abraham, N. & Khan, N. M. A novel focal tversky loss function with improved attention U-Net for lesion segmentation. In Proceedings of the 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI). 683–687 (2019).
Garcea, F., Serra, A., Lamberti, F. & Morra, L. Data augmentation for medical imaging: a systematic literature review. Comput. Biol. Med. 152. https://doi.org/10.1016/j.compbiomed.2022.106391 (2023).
Fan, D. P. et al. PraNet: Parallel reverse attention network for polyp segmentation. In Proceedings of the 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI). 263–273 (2020).
Lou, A. G., Guan, S. Y., Ko, H. & Loew, M. CaraNet: context axial reverse attention network for segmentation of small medical objects. Proc. Spie. 12032. https://doi.org/10.1117/12.2611802 (2022).
Chang, Y. K., Jung, C., Xu, Y. Q. & FinerPCN,. High fidelity point cloud completion network using pointwise convolution. Neurocomputing. 460, 266–276. https://doi.org/10.1016/j.neucom.2021.06.080 (2021).
doi: 10.1016/j.neucom.2021.06.080
Liu, C. L. et al. RB-Net: training highly accurate and efficient binary neural networks with reshaped point-wise convolution and balanced activation. Ieee T Circ. Syst. Vid. 32, 6414–6424. https://doi.org/10.1109/Tcsvt.2022.3166803 (2022).
doi: 10.1109/Tcsvt.2022.3166803
Jang, J. G., Quan, C., Lee, H. D. & Kang, U. Falcon: lightweight and accurate convolution based on depthwise separable convolution. Knowl. Inf. Syst. 65, 2225–2249. https://doi.org/10.1007/s10115-022-01818-x (2023).
doi: 10.1007/s10115-022-01818-x
Tan, M. & Le, Q. EfficientNetV2: smaller models and faster training. In Proceedings of the 38th International Conference on Machine Learning. 10096–10106 (2021).
He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 770–778 (2016).
Ma, S. et al. A mixed visual encoding model based on the larger-scale receptive field for human brain activity. Brain Sci. 12, 1633 (2022).
Wang, Y. X., Wang, J. Y. & Guo, P. Eye-UNet: a UNet-based network with attention mechanism for low-quality human eye image segmentation. Signal. Image Video P. 17, 1097–1103. https://doi.org/10.1007/s11760-022-02316-x (2023).
doi: 10.1007/s11760-022-02316-x