Glaucoma is a chronic optic neuropathy that is one of the leading causes of irreversible blindness worldwide [1], [2]. Hence, it is extremely important to detect it at an early stage with maximum accuracy. The paper presents an innovative approach to an automatic binary classification system based on deep learning for glaucoma detection using the ACRIMA dataset images [3]. The system is based on the transfer learning approach with DenseNet-121 architecture for glaucoma detection. It utilizes contrast enhancement techniques and the extraction of the green channel with optic disc localization for better feature representation. The fine-tuning of the pre-trained DenseNet-121 model is done for binary classification of glaucoma images and normal images. The data was split based on stratified data splitting techniques, resulting in better predictive capabilities for the system with an accuracy of 94.0%, precision of 92.98%, recall of 96.36%, and AUC-ROC of 0.961. The model is better compared to the baseline convolutional neural networks for feature propagation and generalization. Hence, the proposed system can be considered reliable for the diagnosis of glaucoma by ophthalmologists.

This work demonstrates that a transfer-learning approach built on DenseNet121 can reliably identify glaucoma from retinal fundus images with strong classification performance. The model’s architecture — which encourages extensive feature sharing across layers — proved effective at capturing subtle optic-nerve head variations, yielding high accuracy and favorable precision/recall trade-offs on the ACRIMA dataset. The addition of contrast enhancement, refinement of regions of interest, data augmentation, and regularization also helped in obtaining stable training of the model, considering that data is limited in this domain Beyond numeric performance, visualization via class- activation mapping showed that the network’s attention aligns with clinically meaningful regions (primarily the optic disc), supporting the method’s interpretability and potential clinician trust. Low false-negative rates further indicate the approach’s suitability for screening contexts, where missing positive cases carries serious consequences. The pipeline is computationally efficient enough for practical use (with GPU acceleration) and can be integrated into telemedicine or mass-screening workflows to expand access to early glaucoma detection. However, some limitations of the results obtained need to be considered. The results are based on the dataset used for training the network. Therefore, the results need to be validated based on alternative datasets. The deployability of the results is an advantage of the study. The results obtained reveal that the DenseNet121-based framework is an advancement towards the automated screening of glaucoma.

[1]Wong, T.Y., Bressler, N.M., Ting, D.S.W., Artificial intelligence for retinal disease detection, Nature Medicine, 2018. [2]Grabska-Liberek, I., et al., Glaucoma – state of the art and perspectives on treatment, Medical Science Monitor, 2016. [3] Diaz-Pinto, A., Morales, S., Naranjo, V., Köhler, T., Mossi, J.M., Navea, A., CNNs for automatic glaucoma assessment using fundus images, PeerJ Computer Science, 2019. [6]Cheng, J., et al., Superpixel classification based optic disc and optic cup segmentation for glaucoma screening, IEEE Transactions on Medical Imaging, 2013. [7]Screening for Glaucoma in Adults: A Systematic Review for the U.S. Preventive Services Task Force, Agency for Healthcare Research and Quality, 2022. [8]Litjens, G., et al., A survey on deep learning in medical image analysis, Medical Image Analysis, 2017. [9]Prentašić, P., Lončarić, S., Detection of exudates in fundus photographs using deep neural networks, Computer Methods and Programs in Biomedicine, 2016. [10]Zhu, H., Zhang, X., Li, Y., Deep supervised learning methods for medical image classification, IEEE Access, 2020. [11]Simonyan, K., Zisserman, A., Very Deep Convolutional Networks for Large-Scale Image Recognition, ICLR, 2015. [12]Raghavendra, U., Fujita, H., Gudigar, A., Acharya, U.R., Deep learning techniques for glaucoma detection using retinal fundus images, Computers in Biology and Medicine, 2019. [13]Orlando, J.I., Seeböck, P., Bogunović, H., Klimscha, S., Schmidt-Erfurth, U., Deep learning for glaucoma detection using retinal images, Medical Image Analysis, 2019. [14]Raghu, M., et al., Transfusion: Understanding transfer learning for medical imaging, NeurIPS, 2019. [15]Selvaraju, R.R., et al., Grad-CAM: Visual explanations from deep networks via gradient-based localization, ICCV, 2017. [16]Huang, G., Liu, Z., Van der Maaten, L., Weinberger, K.Q., Densely Connected Convolutional Networks, CVPR, 2017. [17]Acharya, U.R., Fujita, H., Lih, O.S., Hagiwara, Y., Tan, J.H., Adam, M., Automated diagnosis of glaucoma using deep learning models, Information Sciences, 2019. [18]Li, J., Liu, L., Xu, J., Automatic glaucoma detection in retinal fundus images using deep convolutional neural networks, IEEE Access, 2019. [19]Roy, S., Bhattacharya, S., Bandyopadhyay, S., Deep learning based medical image analysis: A comprehensive review, Intelligent Systems and Applications, Springer, 2020. [20]Raghavendra, S., Acharya, U.R., Fujita, H., Tan, J.H., Koh, J.E.W., Deep convolution neural network for accurate diagnosis of glaucoma using digital fundus images, Information Sciences, 2018. [21]Thakur, N., Juneja, M., Singh, S., Performance Evaluation of Various Deep Learning Based Approaches for Glaucoma Detection, Journal of Healthcare Engineering, 2022. [22]He, K., Zhang, X., Ren, S., Sun, J., Deep Residual Learning for Image Recognition, CVPR, 2016. [23]Chakravarty, A., Sivaswamy, J., Joint optic disc and cup segmentation for glaucoma detection, IEEE International Symposium on Biomedical Imaging, 2017. [24]Fu, H., Cheng, J., Xu, Y., Wong, D.W.K., Liu, J., Cao, X., Disc-aware ensemble network for glaucoma screening from fundus images, IEEE Transactions on Medical Imaging, 2018. [25]Chen, H., Qi, X., Yu, L., Heng, P., DCAN: Deep contour- aware networks for object instance segmentation from histology images, Medical Image Analysis, 2017. [26]Deng, J., et al., ImageNet: A large-scale hierarchical image database, CVPR, 2009. [27]Tan, M., Le, Q.V., EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks, ICML, 2019. [28]Pathmakumara, H.C., Perera, G., Explainable deep learning for medical imaging: A survey, IEEE Access, 2021. [29]Zhang, Y., Deep learning in glaucoma with fundus photography, Applied and Computational Engineering, 2025. [30]Diaz-Pinto, A., et al., CNNs for automatic glaucoma assessment using fundus images: An extensive validation, Biomedical Engineering Online, 2019. [31]Krishnamoorthi, J., Prasath, V.B.S., Selvakumar, J., Deep learning for automated glaucoma screening from retinal fundus images, International Journal of Advanced Computer Science and Applications, 2019. [32]Al-Bander, B., Williams, B.M., Al-Nuaimy, W., Al-Taee, M.A., Pratt, H., Zheng, Y., Automated glaucoma screening using deep learning and optic disc segmentation, IEEE Journal of Biomedical and Health Informatics, 2019. [33] Fu, H., Cheng, J., Xu, Y., Wong, D.W.K., Liu, J., Cao, X., REFUGE challenge: A unified framework for evaluating automated glaucoma assessment methods, Medical Image Analysis, 2020. [34]Li, L., Xu, M., Wang, X., Jiang, H., Deep learning for detecting retinal diseases from fundus photographs, IEEE Access, 2020. [35]Chen, J., Zhang, Y., Li, W., Fusion neural networks for medical image classification, Lecture Notes in Computer Science, Springer, 2021. [36]Ferro, P., Costa, M., Silva, A., Computational techniques for medical image analysis, Springer, 2020. [37]Toka, D., Kirci, M., Deep learning based bone age assessment using convolutional neural networks, Springer, 2021. [38]Gupta, B.B., Nedjah, N., Safety, security and reliability of robotic systems, Springer, 2021. [39]Sethi, S., Sharma, M., Gupta, A., Artificial intelligence methods in medical image analysis, Journal of Artificial Intelligence Research, 2021. [40]Sambasivarao, K.V., Bhima, A.S.R.D., Artificial intelligence techniques in medical imaging, Springer, 2021. [41]Mouhafid, L., Bencherif, M., Taleb-Ahmed, A., Efficient neural models for medical image classification, IEEE Access, 2021. [42]Ting, D.S.W., Cheung, C.Y.L., Wong, T.Y., Development and validation of a deep learning system for disease detection in retinal images, Nature Medicine, 2018. [43]Li, A., Zhao, Y., Zhang, X., Transfer learning for medical image classification using DenseNet, IEEE Access, 2020. [44]Fawcett, T., An introduction to ROC analysis, Pattern Recognition Letters, 2006. [45]Gómez-Valverde, J., Antón, A., Fatti, G., Liefers, B., Herranz, A., Santos, A., Automatic glaucoma classification using color fundus images based on convolutional neural networks and transfer learning, IEEE Access, 2019. [46]Li, Z., et al., Efficacy of deep learning system for detecting glaucoma, JAMA Ophthalmology, 2018. [47]Zuiderveld, P., Contrast Limited Adaptive Histogram Equalization, Graphics Gems IV, 1994. [48]Kingma, D., Ba, J., Adam: A Method for Stochastic Optimization, ICLR, 2015. [49]Duda, R., Hart, P., Use of the Hough transformation to detect lines and curves in pictures, Communications of the ACM, 1972. [50]Fumero, F., et al., RIM-ONE DL: A Retinal Image Database for Deep Learning-Based Glaucoma Assessment, 2020. [51]Kohavi, R., A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection, IJCAI, 1995. [52]Fu, A., Cheng, J., Xu, Y., Wong, D., Liu, J., Wong, T.Y., Joint optic disc and cup segmentation based on multi-label deep network, IEEE Transactions on Medical Imaging, 2018. [53]Tham, K.A., et al., Automated glaucoma detection from fundus images using deep learning techniques, IEEE Access, 2020. [54]Al-Bander, M., Williams, B., Al-Nuaimy, W., Al-Taee, M., Pratt, H., Zheng, Y., Dense fully convolutional segmentation of optic disc and cup, Symmetry, 2018. [55] Orlando, J., Prokofyeva, E., Blaschko, M.B., Conditional random field model for blood vessel segmentation in fundus images, IEEE Transactions on Biomedical Engineering, 2017. [55]Ling, X.C., et al., Deep learning in glaucoma detection and progression prediction: A systematic review and meta- analysis, Biomedicines, 2025. [56]Mouhafid, M., Zhou, Y., Shan, C., Xiao, Z., Modular convolution–involution cascade architecture for glaucoma screening, PeerJ Computer Science, 2025. [57]Latha, I.G., Aruna Priya, P., Glaucoma detection and severity diagnosis using dual CNN architectures, International Journal of Image Graphics and Signal Processing, 2024. [58]Singh, A.K., Iterative data augmentation for improving deep learning performance in medical image classification, Multimedia Tools and Applications, 2022.

© 2025 International Journal of Engineering and Techniques (IJET).

Digital Object Identifier (DOI)

IJET assigns a unique DOI to every accepted article via Zenodo for persistent linking and citation. Your article’s DOI: https://doi.org/{{doi}}

Open DOI About Zenodo DOI

Close