{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,12]],"date-time":"2026-05-12T06:52:57Z","timestamp":1778568777327,"version":"3.51.4"},"reference-count":50,"publisher":"MDPI AG","issue":"8","license":[{"start":{"date-parts":[[2025,8,15]],"date-time":"2025-08-15T00:00:00Z","timestamp":1755216000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"University Kuala Lumpur (UniKL), Malaysia"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J. Imaging"],"abstract":"<jats:p>Timely, balanced, and transparent detection of retinal diseases is essential to avert irreversible vision loss; however, current deep learning screeners are hampered by class imbalance, large models, and opaque reasoning. This paper presents a lightweight attention-augmented convolutional neural network (CNN) that addresses all three barriers. The network combines depthwise separable convolutions, squeeze-and-excitation, and global-context attention, and it incorporates gradient-based class activation mapping (Grad-CAM) and Grad-CAM++ to ensure that every decision is accompanied by pixel-level evidence. A 5335-image ten-class color-fundus dataset from Bangladeshi clinics, which was severely skewed (17\u20131509 images per class), was equalized using a synthetic minority oversampling technique (SMOTE) and task-specific augmentations. Images were resized to 150\u00d7150 px and split 70:15:15. The training used the adaptive moment estimation (Adam) optimizer (initial learning rate of 1\u00d710\u22124, reduce-on-plateau, early stopping), \u21132 regularization, and dual dropout. The 16.6 M parameter network converged in fewer than 50 epochs on a mid-range graphics processing unit (GPU) and reached 87.9% test accuracy, a macro-precision of 0.882, a macro-recall of 0.879, and a macro-F1-score of 0.880, reducing the error by 58% relative to the best ImageNet backbone (Inception-V3, 40.4% accuracy). Eight disorders recorded true-positive rates above 95%; macular scar and central serous chorioretinopathy attained F1-scores of 0.77 and 0.89, respectively. Saliency maps consistently highlighted optic disc margins, subretinal fluid, and other hallmarks. Targeted class re-balancing, lightweight attention, and integrated explainability, therefore, deliver accurate, transparent, and deployable retinal screening suitable for point-of-care ophthalmic triage on resource-limited hardware.<\/jats:p>","DOI":"10.3390\/jimaging11080275","type":"journal-article","created":{"date-parts":[[2025,8,15]],"date-time":"2025-08-15T14:24:28Z","timestamp":1755267868000},"page":"275","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["A Lightweight CNN for Multiclass Retinal Disease Screening with Explainable AI"],"prefix":"10.3390","volume":"11","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-2244-2328","authenticated-orcid":false,"given":"Arjun Kumar Bose","family":"Arnob","sequence":"first","affiliation":[{"name":"Department of Computer Science, American International University-Bangladesh, Dhaka 1229, Bangladesh"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3741-6651","authenticated-orcid":false,"given":"Muhammad Hasibur Rashid","family":"Chayon","sequence":"additional","affiliation":[{"name":"Department of Computer Science, American International University-Bangladesh, Dhaka 1229, Bangladesh"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2625-2348","authenticated-orcid":false,"given":"Fahmid","family":"Al Farid","sequence":"additional","affiliation":[{"name":"Faculty of Computer Science and Informatics, Berlin School of Business and Innovation, 12043 Berlin, Germany"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3791-2436","authenticated-orcid":false,"given":"Mohd Nizam","family":"Husen","sequence":"additional","affiliation":[{"name":"Malaysian Institute of Information Technology, Universiti Kuala Lumpur, Kuala Lumpur 50250, Malaysia"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-0007-7294","authenticated-orcid":false,"given":"Firoz","family":"Ahmed","sequence":"additional","affiliation":[{"name":"Department of Computer Science, American International University-Bangladesh, Dhaka 1229, Bangladesh"}]}],"member":"1968","published-online":{"date-parts":[[2025,8,15]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"17","DOI":"10.1038\/s41433-022-02045-y","article-title":"The Peter Watson Memorial Lecture \u201cVision for the World\u201d","volume":"37","author":"Taylor","year":"2023","journal-title":"Eye"},{"key":"ref_2","first-page":"493","article-title":"Updates on the Epidemiology of Age-Related Macular Degeneration","volume":"6","author":"Jonas","year":"2017","journal-title":"Asia-Pac. J. Ophthalmol."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"369","DOI":"10.1038\/s41433-021-01684-x","article-title":"Eye Disease and Mortality, Cognition, Disease, and Modifiable Risk Factors: An Umbrella Review of Meta-Analyses of Observational Studies","volume":"36","author":"Trott","year":"2022","journal-title":"Eye"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"98","DOI":"10.1097\/ICU.0000000000000340","article-title":"The Global State of Cataract Blindness","volume":"28","author":"Lee","year":"2017","journal-title":"Curr. Opin. Ophthalmol."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"1123","DOI":"10.1001\/jamaophthalmol.2024.4401","article-title":"Economic Burden of Late-Stage Age-Related Macular Degeneration in Bulgaria, Germany, and the US","volume":"142","author":"Paudel","year":"2024","journal-title":"JAMA Ophthalmol."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"587","DOI":"10.1007\/s00417-019-04484-y","article-title":"Socioeconomic Disparities in the Global Burden of Glaucoma: An Analysis of Trends from 1990 to 2016","volume":"258","author":"Wu","year":"2020","journal-title":"Graefe\u2019s Arch. Clin. Exp. Ophthalmol."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"1295","DOI":"10.1001\/jamaophthalmol.2017.3449","article-title":"Association of Socioeconomics With Prevalence of Visual Impairment and Blindness","volume":"135","author":"Wang","year":"2017","journal-title":"JAMA Ophthalmol."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"e138","DOI":"10.1111\/aos.14198","article-title":"Socio-Economic Disparity in Global Burden of near Vision Loss: An Analysis for 2017 with Time Trends since 1990","volume":"98","author":"Wang","year":"2020","journal-title":"Acta Ophthalmol."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Ejaz, S., Baig, R., Ashraf, Z., Alnfiai, M.M., Alnahari, M.M., and Alotaibi, R.M. (2024). A Deep Learning Framework for the Early Detection of Multi-Retinal Diseases. PLoS ONE, 19.","DOI":"10.1371\/journal.pone.0307317"},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Ghidoni, S. (2024). The Role of Deep Learning in the Diagnosis of Ocular Diseases. Acta Ophthalmol., 102.","DOI":"10.1111\/aos.16547"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1007\/s41666-024-00182-5","article-title":"A Low Complexity Efficient Deep Learning Model for Automated Retinal Disease Diagnosis","volume":"9","author":"Chowa","year":"2025","journal-title":"J. Healthc. Inform. Res."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"171221","DOI":"10.1109\/ACCESS.2024.3496565","article-title":"Advancements in Deep Learning for Automated Diagnosis of Ophthalmic Diseases: A Comprehensive Review","volume":"12","author":"Dash","year":"2024","journal-title":"IEEE Access"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"57796","DOI":"10.1109\/ACCESS.2022.3178372","article-title":"A Comprehensive Review of Deep Learning Strategies in Retinal Disease Diagnosis Using Fundus Images","volume":"10","author":"Goutam","year":"2022","journal-title":"IEEE Access"},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Vidivelli, S., Padmakumari, P., Parthiban, C., DharunBalaji, A., Manikandan, R., and Gandomi, A.H. (2025). Optimising Deep Learning Models for Ophthalmological Disorder Classification. Sci. Rep., 15.","DOI":"10.1038\/s41598-024-75867-3"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"29","DOI":"10.1167\/tvst.12.1.29","article-title":"Deep Learning-Based System for Disease Screening and Pathologic Region Detection From Optical Coherence Tomography Images","volume":"12","author":"Chen","year":"2023","journal-title":"Transl. Vis. Sci. Technol."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"16","DOI":"10.1167\/tvst.13.2.16","article-title":"Deep Learning and Machine Learning Algorithms for Retinal Image Analysis in Neurodegenerative Disease: Systematic Review of Datasets and Models","volume":"13","author":"Bahr","year":"2024","journal-title":"Transl. Vis. Sci. Technol."},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Elreedy, D., and Atiya, A.F. (2019, January 12\u201314). A Novel Distribution Analysis for SMOTE Oversampling Method in Handling Class Imbalance. Proceedings of the Computational Science\u2014ICCS 2019, Faro, Portugal.","DOI":"10.1007\/978-3-030-22744-9_18"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"117023","DOI":"10.1016\/j.eswa.2022.117023","article-title":"A Novel SMOTE-based Resampling Technique Trough Noise Detection and the Boosting Procedure","volume":"200","author":"Cengiz","year":"2022","journal-title":"Expert Syst. Appl."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"158742","DOI":"10.1109\/ACCESS.2024.3480848","article-title":"FADA-SMOTE-Ms: Fuzzy Adaptative Smote-Based Methods","volume":"12","author":"Mohammed","year":"2024","journal-title":"IEEE Access"},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Bunkhumpornpat, C., Sinapiromsaran, K., and Lursinsap, C. (2009). Safe-Level-SMOTE: Safe-Level-Synthetic Minority Over-Sampling TEchnique for Handling the Class Imbalanced Problem. Advances in Knowledge Discovery and Data Mining, Springer.","DOI":"10.1007\/978-3-642-01307-2_43"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"74045","DOI":"10.1109\/ACCESS.2022.3187699","article-title":"Synthetic Minority Oversampling Technique Based on Adaptive Local Mean Vectors and Improved Differential Evolution","volume":"10","author":"Zhang","year":"2022","journal-title":"IEEE Access"},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"2456","DOI":"10.1109\/TBME.2020.3043215","article-title":"Robust and Interpretable Convolutional Neural Networks to Detect Glaucoma in Optical Coherence Tomography Images","volume":"68","author":"Thakoor","year":"2021","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"84785","DOI":"10.1109\/ACCESS.2024.3415617","article-title":"Deep Learning in Automatic Diabetic Retinopathy Detection and Grading Systems: A Comprehensive Survey and Comparison of Methods","volume":"12","author":"Abushawish","year":"2024","journal-title":"IEEE Access"},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"417","DOI":"10.1136\/bjo-2022-322183","article-title":"Ensemble of Deep Convolutional Neural Networks Is More Accurate and Reliable than Board-Certified Ophthalmologists at Detecting Multiple Diseases in Retinal Fundus Photographs","volume":"108","author":"Pandey","year":"2024","journal-title":"Br. J. Ophthalmol."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"3406","DOI":"10.1118\/1.4955912","article-title":"SU-F-J-04: Automated Detection of Diabetic Retinopathy Using Deep Convolutional Neural Networks","volume":"43","author":"Xu","year":"2016","journal-title":"Med. Phys."},{"key":"ref_26","first-page":"1","article-title":"CA-Net: A Novel Cascaded Attention-Based Network for Multistage Glaucoma Classification Using Fundus Images","volume":"72","author":"Das","year":"2023","journal-title":"IEEE Trans. Instrum. Meas."},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Moya-S\u00e1nchez, E.U., S\u00e1nchez, A., Zapata, M., Moreno, J., Garcia-Gasulla, D., Parr\u00e9s, F., Ayguad\u00e9, E., Labarta, J., and Cort\u00e9s, U. (2019). Data Augmentation for Deep Learning of Non-mydriatic Screening Retinal Fundus Images. Supercomputing, Springer.","DOI":"10.1007\/978-3-030-10448-1_18"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"12561","DOI":"10.1007\/s10462-023-10453-z","article-title":"Medical Image Data Augmentation: Techniques, Comparisons and Interpretations","volume":"56","author":"Goceri","year":"2023","journal-title":"Artif. Intell. Rev."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"111636","DOI":"10.1109\/ACCESS.2022.3215126","article-title":"Impact of Generative Modeling for Fundus Image Augmentation with Improved and Degraded Quality in the Classification of Glaucoma","volume":"10","author":"Leonardo","year":"2022","journal-title":"IEEE Access"},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Mounsaveng, S., Laradji, I., Ayed, I.B., V\u00e1zquez, D., and Pedersoli, M. (2021, January 3\u20138). Learning Data Augmentation with Online Bilevel Optimization for Image Classification. Proceedings of the 2021 IEEE Winter Conference on Applications of Computer Vision (WACV), Waikoloa, HI, USA.","DOI":"10.1109\/WACV48630.2021.00173"},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Zhang, C., Tavanapong, W., Wong, J., de Groen, P.C., and Oh, J. (2017). Real Data Augmentation for Medical Image Classification. Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis, Springer.","DOI":"10.1007\/978-3-319-67534-3_8"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"101095","DOI":"10.1016\/j.xcrm.2023.101095","article-title":"Artificial Intelligence in Ophthalmology: The Path to the Real-World Clinic","volume":"4","author":"Li","year":"2023","journal-title":"Cell Rep. Med."},{"key":"ref_33","first-page":"187","article-title":"Artificial Intelligence and Optical Coherence Tomography Imaging","volume":"8","author":"Kapoor","year":"2019","journal-title":"Asia-Pac. J. Ophthalmol."},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Ji, Y., Liu, S., Hong, X., Lu, Y., Wu, X., Li, K., Li, K., and Liu, Y. (2022). Advances in Artificial Intelligence Applications for Ocular Surface Diseases Diagnosis. Front. Cell Dev. Biol., 10.","DOI":"10.3389\/fcell.2022.1107689"},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"38","DOI":"10.1186\/s40662-024-00405-1","article-title":"Advances and Prospects of Multi-Modal Ophthalmic Artificial Intelligence Based on Deep Learning: A Review","volume":"11","author":"Wang","year":"2024","journal-title":"Eye Vis."},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"110979","DOI":"10.1016\/j.dib.2024.110979","article-title":"A Dataset of Color Fundus Images for the Detection and Classification of Eye Diseases","volume":"57","author":"Sharmin","year":"2024","journal-title":"Data Brief"},{"key":"ref_37","doi-asserted-by":"crossref","unstructured":"He, K., Zhang, X., Ren, S., and Sun, J. (2015). Deep Residual Learning for Image Recognition. arXiv.","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref_38","doi-asserted-by":"crossref","unstructured":"Hu, J., Shen, L., and Sun, G. (2018, January 18\u201323). Squeeze-and-Excitation Networks. Proceedings of the 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.","DOI":"10.1109\/CVPR.2018.00745"},{"key":"ref_39","unstructured":"Ioffe, S., and Szegedy, C. (2015, January 7\u20139). Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. Proceedings of the 32nd International Conference on International Conference on Machine Learning\u2014Volume 37, ICML\u201915, Lille, France."},{"key":"ref_40","unstructured":"Nair, V., and Hinton, G.E. (2010, January 21\u201324). Rectified Linear Units Improve Restricted Boltzmann Machines. Proceedings of the 27th International Conference on International Conference on Machine Learning, ICML\u201910, Madison, WI, USA."},{"key":"ref_41","unstructured":"Lin, M., Chen, Q., and Yan, S. (2014). Network in Network. arXiv."},{"key":"ref_42","doi-asserted-by":"crossref","unstructured":"Chollet, F. (2017, January 21\u201326). Xception: Deep Learning with Depthwise Separable Convolutions. Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA.","DOI":"10.1109\/CVPR.2017.195"},{"key":"ref_43","unstructured":"Howard, A.G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., and Adam, H. (2017). MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv."},{"key":"ref_44","doi-asserted-by":"crossref","unstructured":"Cao, Y., Xu, J., Lin, S., Wei, F., and Hu, H. (2019, January 27\u201328). GCNet: Non-Local Networks Meet Squeeze-Excitation Networks and Beyond. Proceedings of the 2019 IEEE\/CVF International Conference on Computer Vision Workshop (ICCVW), Seoul, Republic of Korea.","DOI":"10.1109\/ICCVW.2019.00246"},{"key":"ref_45","doi-asserted-by":"crossref","unstructured":"Yacouby, R., and Axman, D. (2020, January 20). Probabilistic Extension of Precision, Recall, and F1 Score for More Thorough Evaluation of Classification Models. Proceedings of the First Workshop on Evaluation and Comparison of NLP Systems, Online.","DOI":"10.18653\/v1\/2020.eval4nlp-1.9"},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"3351","DOI":"10.1109\/JBHI.2020.3011805","article-title":"How to Extract More Information With Less Burden: Fundus Image Classification and Retinal Disease Localization With Ophthalmologist Intervention","volume":"24","author":"Meng","year":"2020","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_47","doi-asserted-by":"crossref","first-page":"127816","DOI":"10.1016\/j.neucom.2024.127816","article-title":"Retinal Disease Diagnosis with Unsupervised Grad-CAM Guided Contrastive Learning","volume":"593","author":"Zhao","year":"2024","journal-title":"Neurocomputing"},{"key":"ref_48","doi-asserted-by":"crossref","unstructured":"Kim, M., Park, H.m., Zuallaert, J., Janssens, O., Hoecke, S.V., and Neve, W.D. (2018, January 3\u20136). Computer-Aided Diagnosis and Localization of Glaucoma Using Deep Learning. Proceedings of the 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Madrid, Spain.","DOI":"10.1109\/BIBM.2018.8621168"},{"key":"ref_49","doi-asserted-by":"crossref","first-page":"29943","DOI":"10.1007\/s11042-022-14165-4","article-title":"Detection of Diabetic Retinopathy Using Convolutional Neural Networks for Feature Extraction and Classification (DRFEC)","volume":"82","author":"Das","year":"2023","journal-title":"Multimed. Tools Appl."},{"key":"ref_50","doi-asserted-by":"crossref","unstructured":"Saito, M., Mitamura, M., Kimura, M., Ito, Y., Endo, H., Katsuta, S., Kase, M., and Ishida, S. (2024). Grad-CAM-Based Investigation into Acute-Stage Fluorescein Angiography Images to Predict Long-Term Visual Prognosis of Branch Retinal Vein Occlusion. J. Clin. Med., 13.","DOI":"10.3390\/jcm13175271"}],"container-title":["Journal of Imaging"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2313-433X\/11\/8\/275\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T18:28:40Z","timestamp":1760034520000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2313-433X\/11\/8\/275"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8,15]]},"references-count":50,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2025,8]]}},"alternative-id":["jimaging11080275"],"URL":"https:\/\/doi.org\/10.3390\/jimaging11080275","relation":{},"ISSN":["2313-433X"],"issn-type":[{"value":"2313-433X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,8,15]]}}}