Optimized Neuro-Fuzzy Classification of Glaucoma from Fundus Images Using Modified Cuckoo Search and Gabor Features
Asma Naseri
1
(
Department of Electrical and Biomedical Engineering, University of Bojnord, Bojnord, Iran.
)
Ali Khazaee
2
(
Department of Electrical and Biomedical Engineering, Assistant Professor of Electrical Engineering, University of Bojnord, Bojnord, Iran.
)
Keywords: Glaucoma Detection, Fundus Image Classification, Gabor Filter, Adaptive Neuro-Fuzzy Inference System (ANFIS), ANFIS Optimization, Modified Cuckoo Search (MCS), Clinical Screening Support.,
Abstract :
This study proposes a hybrid framework for automated glaucoma detection and classification from retinal fundus images, leveraging an Adaptive Neuro-Fuzzy Inference System (ANFIS) optimized via a novel Modified Cuckoo Search (MCS) algorithm, designed to enhance convergence and accuracy in fuzzy rule training. Fundus images from the publicly available RIGA dataset were preprocessed to isolate the optic disc and cup using a Gabor wavelet-based filtering method applied to the green channel, which offers superior vascular contrast. Morphological parameters, specifically the Cup-to-Disc Ratio (CDR) and Rim-to-Disc Ratio (RDR), were extracted to form a discriminative feature set. An ANFIS model was then trained to classify images as glaucomatous or healthy. Optimization of ANFIS membership function parameters was performed using both Particle Swarm Optimization (PSO) and the proposed MCS algorithm for comparison. Model performance was evaluated using Root Mean Square Error (RMSE), Mean Squared Error (MSE), regression coefficient (R), and area under the ROC curve (AUC). The MCSoptimized ANFIS outperformed the PSO-based model, achieving a lower RMSE of 0.0009963 (vs. 0.001723) and a higher test-phase regression coefficient of R = 0.9832 (vs. R = 0.4811). This corresponds to a 42.15% reduction in RMSE and a 104.4% relative improvement in R, demonstrating the significance of the MCS optimizer in enhancing model accuracy. The model demonstrated a high AUC of 0.90 for vessel segmentation and maintained robust classification accuracy across a 70 : 30 train-test split on 195 annotated fundus images. The proposed MCS-optimized ANFIS framework offers a reliable and interpretable solution for early glaucoma detection. Its superior performance in both segmentation and classification tasks highlights its potential for integration into clinical decision-support systems for ophthalmic screening.
[1] Blumberg MJ, Varikuti VN, Weiner A. Real-world comparison between the Tonopen and Goldmann applanation tonometry in a university glaucoma clinic. International Ophthalmology. 2021; 41: 18151825. DOI: https://doi.org/10.1007/s10792-021-01742-z
[2] Xu J, Li R, Xu H, Yang Y, Zhang S, Ren TL. Recent progress of continuous intraocular pressure monitoring. Nano Select. 2022; 3(1): 119. DOI: https://doi.org/10.1002/nano.202100137
[3] Ilesanmi AE, Ilesanmi T, Gbotoso GA. A systematic review of retinal fundus image segmentation and classification methods using convolutional neural networks. Healthcare Analytics. 2023; 4: 100261. DOI: https://doi.org/10.1016/j.health.2023.100261
[4] Panda R, N.B Puhan, Panda G. Robust and accurate optic disk localization using vessel symmetry line measure in fundus images. Biocybernetics and Biomedical Engineering. 2017; 37(3): 466-476. DOI: https://doi.org/10.1016/j.bbe.2017.05.008
[5] Khan SI, Choubey SB, Choubey A, Bhatt A, Naishadhkumar PV, Basha MM. Automated glaucoma detection from fundus images using wavelet-based denoising and machine learning. Concurrent Engineering. 2021; 30(1): 103115. DOI: https://doi.org/10.1177/1063293x211026620
[6] Suot D, Alonso-Caneiro D, Ksieniewicz P, Krzyzanowska-Berkowska P, Iskander DR. Glaucoma classification based on scanning laser ophthalmoscopic images using a deep learning ensemble method. PLOS ONE. 2021; 16(6): e0252339. DOI: https://doi.org/10.1371/journal.pone.0252339
[7] Singh LK, Khanna M, Thawkar S. A novel hybrid robust architecture for automatic screening of glaucoma using fundus photos, built on feature selection and machine learningnature driven computing. Expert Systems. 2022; 39(10): e13069. DOI: https://doi.org/10.1111/exsy.13069
[8] Pathan S, Kumar P, Pai RM, Bhandary SV. An automated classification framework for glaucoma detection in fundus images using ensemble of dynamic selection methods. Progress in Artificial Intelligence. 2023; 12(3): 287301. DOI: https://doi.org/10.1007/s13748-023-00304-x
[9] Feng Y, Li Z, Yang D, Hu H, Guo H, Liu H. Polarformer: Optic disc and cup segmentation using a hybrid CNN-transformer and polar transformation. Applied Sciences. 2023; 13(1): 541. DOI: https://doi.org/10.3390/app13010541
[10] Joshi S, Partibane B, Hatamleh WA, Tarazi H, Yadav CS, Krah D. Glaucoma detection using image processing and supervised learning for classification. Journal of Healthcare Engineering. 2022; 2022: 2988262. DOI: https://doi.org/10.1155/2022/2988262
[11] Ma F, Li S, Wang S, Guo Y, Wu F, Meng J, Dai C. Deep-learning segmentation method for optical coherence tomography angiography in ophthalmology. Journal of Biophotonics. 2024; 17(2): e202300321. DOI: https://doi.org/10.1002/jbio.202300321
[12] Maninis KK, Pont-Tuset J, Arbelez P, Van Gool L. Deep retinal image understanding. In: Ourselin, S., Joskowicz, L., Sabuncu, M., Unal, G., Wells, W. (eds.) Medical Image Computing and Computer-Assisted Intervention MICCAI 2016: Proceedings Part II, 1721 Oct 2016, Athens, Greece. Cham: Springer; 2016. p.140148. DOI: https://doi.org/10.1007/978-3-319-46723-8 17
[13] Zilly J, Buhmann JM, Mahapatra D. Glaucoma detection using entropy sampling and ensemble learning for automatic optic cup and disc segmentation. Computerized Medical Imaging and Graphics. 2017; 55: 2841. DOI: https://doi.org/10.1016/j.compmedimag.2016.07.012
[14] Tan JH, Acharya UR, Bhandary SV, Chua KC, Sivaprasad S. Segmentation of optic disc, fovea and retinal vasculature using a single convolutional neural network. Journal of Computational Science. 2017; 20: 7079. DOI: https://doi.org/10.1016/j.jocs.2017.02.006
[15] Mayee MK, Khanam MH. Simplifying glaucoma diagnosis with U-Net on retinal images. Traitement du Signal. 2024; 41(4): 21592168. DOI: https://doi.org/10.18280/ts.410444
[16] Meedeniya D, Shyamalee T, Lim G, Yogarajah P. Glaucoma identification with retinal fundus images using deep learning: Systematic review. Informatics in Medicine Unlocked. 2025; 56: 101644. DOI: https://doi.org/10.1016/j.imu.2025.101644
[17] Jang JSR. ANFIS: adaptive-network-based fuzzy inference system. IEEE Transactions on Systems, Man and Cybernetics. 1993; 23(3): 665685. DOI: https://doi.org/10.1109/21.256541.
[18] Shyamalee T, Meedeniya D. Attention U-Net for glaucoma identification using fundus image segmentation. In: Proceedings of the 2022 International Conference on Decision Aid Sciences and Applications (DASA); 2325 Mar 2022, Chiangrai, Thailand. Piscataway, NJ: IEEE; 2022. p.610. DOI: https://doi.org/10.1109/DASA54658.2022.9765303
[19] Aquino A, Gegndez-Arias ME, Marn D. Detecting the optic disc boundary in digital fundus images using morphological, edge detection, and feature extraction techniques. IEEE Transactions on Medical Imaging. 2010; 29(11): 18601869. DOI: https://doi.org/10.1109/TMI.2010.2053042
[20] Lowell J, Hunter A, Steel D, Basu A, Ryder R, Fletcher E, Kennedy L. Optic nerve head segmentation. IEEE Transactions on Medical Imaging. 2004; 23(2): 256264. DOI:
https://doi.org/10.1109/TMI.2003.823261
[21] Lu S. Accurate and efficient optic disc detection and segmentation by a circular transformation. IEEE Transactions on Medical Imaging. 2011; 30(12): 21262133. DOI:
https://doi.org/10.1109/TMI.2011.2164261
[22] Shyamalee T, Meedeniya D. CNN based fundus images classification for glaucoma identification. In: Proceedings of the 2022 2nd International Conference on Advanced Research in Computing (ICARC); 2324 Feb 2022. Piscataway, NJ: IEEE; 2022. p.200205. DOI: https://doi.org/10.1109/ICARC54489.2022.9754171
[23] Diaz-Pinto A, Morales S, Naranjo V, Khler T, Mossi JM, Navea A. CNNs for automatic glaucoma assessment using fundus images: an extensive validation. BioMedical Engineering OnLine. 2019; 18(29). DOI: https://doi.org/10.1186/s12938-019-0649-y
[24] Asaoka R, Tanito M, Shibata N, Mitsuhashi K, Nakahara K, Fujino Y, et al. Validation of a deep learning model to screen for glaucoma using images from different fundus cameras and data augmentation. Ophthalmology Glaucoma. 2019; 2(4): 224231. DOI: https://doi.org/10.1016/j.ogla.2019.03.008
[25] Shyamalee T, Meedeniya D. Glaucoma Detection with Retinal Fundus Images Using Segmentation and Classification. Machine Intelligence Research. 2022; 19(6): 563580. DOI: https://doi.org/10.1007/s11633-022-1354-z
[26] Cheng J, Liu J, Xu Y, Yin F, Wong DWK, Tan NM, Tao D, Cheng CY, Aung T, Wong TY. Superpixel classification based optic disc and optic cup segmentation for glaucoma screening. IEEE Transactions on Medical Imaging. 2013; 32(6): 10191032. DOI: https://doi.org/10.1109/TMI.2013.2247770
[27] Fu H, Cheng J, Xu Y, Zhang C, Wong DWK, Liu J, Cao X. Disc-aware ensemble network for glaucoma screening from fundus image. IEEE Transactions on Medical Imaging. 2018; 37(11): 24932501. DOI: https://doi.org/10.1109/TMI.2018.2837012
[28] Almazroa A, Alodhayb S, Osman E, Ramadan E, Hummadi M, Dlaim M, Alkatee M, Raahemifar K, Lakshminarayanan V. Retinal fundus images for glaucoma analysis: the RIGA dataset. In: Medical Imaging 2018: Imaging Informatics for Healthcare, Research, and Applications. Proc. SPIE 10579; 2018 Mar 6. p.105790B. DOI: https://doi.org/10.1117/12.2293584
[29] Martinez-Perez ME, Witt N, Parker KH, Hughes AD, Thom SA. Automatic optic disc detection in colour fundus images by means of multispectral analysis and information content. PeerJ. 2019; 7: e7119. DOI: https://doi.org/10.7717/peerj.7119.
[30] Yang XS, Suash D. Cuckoo search via Lvy flights. In: 2009 World Congress on Nature & Biologically Inspired Computing (NaBIC); 2009 Dec 911. pp. 210214. DOI:
https://doi.org/10.1109/NABIC.2009.5393690.
[31] Soares JVB, Leandro JJG, Cesar RM Jr, Jelinek HF, Cree MJ. Retinal vessel segmentation using the 2-D Gabor wavelet and supervised classification. IEEE Transactions on Medical Imaging. 2006; 25(9): 12141222. DOI: https://doi.org/10.1109/TMI.2006.879967.
[32] Kumar JRH, Seelamantula CS, Kamath YS, Jampala R. Rim-to-disc ratio outperforms cup-to-disc ratio for glaucoma prescreening. Scientific Reports. 2019; 9: 7099. DOI: https://doi.org/10.1038/s41598-019- 43385-2.
[33] Shyamalee T, Meedeniya D, Lim G, Karunarathne M. Automated tool support for glaucoma identification with explainability using fundus images. IEEE Access. 2024; 12: 1729017307. DOI: https://doi.org/10.1109/ACCESS.2024.3359698.
[34] Das D, Nayak DR. GS-Net: Global self-attention guided CNN for multi-stage glaucoma classification. In: 2023 IEEE International Conference on Image Processing (ICIP); 811 Oct 2023. p.34543458. DOI: https://doi.org/10.1109/ICIP49359.2023.10222689.
[35] Li Y, Carrillo-Perez F, Owaifeer AM, Alawad M, Gevaert O. Enhancing glaucoma detection through supervised pre-training with intermediate phenotypes: A multi-institutional study. medRxiv [Preprint] 2025. DOI: https://doi.org/10.1101/2025.04.22.25326210.
[36] Luo Y, Shi M, Tian Y, Elze T, Wang M. Harvard glaucoma detection and progression: A multimodal multitask dataset and generalization-reinforced semi-supervised learning. In: 2023 IEEE/CVF International Conference on Computer Vision (ICCV); 1-6 Oct 2023. p.2041420425. DOI: https://doi.org/10.1109/ICCV51070.2023.01872.