Ramon Pires1, Sandra Avila2, Jacques Wainer3, Eduardo Valle4, Michael D Abramoff5, Anderson Rocha6. 1. Institute of Computing, University of Campinas (Unicamp), Campinas 13083-852, Brazil. Electronic address: pires.ramon@ic.unicamp.br. 2. Institute of Computing, University of Campinas (Unicamp), Campinas 13083-852, Brazil. Electronic address: sandra@ic.unicamp.br. 3. Institute of Computing, University of Campinas (Unicamp), Campinas 13083-852, Brazil. Electronic address: wainer@ic.unicamp.br. 4. School of Electrical and Computing Engineering, University of Campinas (Unicamp), Campinas 13083-852, Brazil. Electronic address: dovalle@dca.fee.unicamp.br. 5. Stephen R. Wynn Institute for Vision Research, the Department of Electrical and Computer Engineering, the Department of Biomedical Engineering, the University of Iowa, Iowa City, IA 52242, USA; VA Medical Center, Iowa City, IA 52246, USA; IDx LLC, Iowa City, IA, USA. Electronic address: michael-abramoff@uiowa.edu. 6. Institute of Computing, University of Campinas (Unicamp), Campinas 13083-852, Brazil. Electronic address: anderson.rocha@ic.unicamp.br.
Abstract
Prior art on automated screening of diabetic retinopathy and direct referral decision shows promising performance; yet most methods build upon complex hand-crafted features whose performance often fails to generalize. OBJECTIVE: We investigate data-driven approaches that extract powerful abstract representations directly from retinal images to provide a reliable referable diabetic retinopathy detector. METHODS: We gradually build the solution based on convolutional neural networks, adding data augmentation, multi-resolution training, robust feature-extraction augmentation, and a patient-basis analysis, testing the effectiveness of each improvement. RESULTS: The proposed method achieved an area under the ROC curve of 98.2% (95% CI: 97.4-98.9%) under a strict cross-dataset protocol designed to test the ability to generalize - training on the Kaggle competition dataset and testing using the Messidor-2 dataset. With a 5 × 2-fold cross-validation protocol, similar results are achieved for Messidor-2 and DR2 datasets, reducing the classification error by over 44% when compared to most published studies in existing literature. CONCLUSION: Additional boost strategies can improve performance substantially, but it is important to evaluate whether the additional (computation- and implementation-) complexity of each improvement is worth its benefits. We also corroborate that novel families of data-driven methods are the state of the art for diabetic retinopathy screening. SIGNIFICANCE: By learning powerful discriminative patterns directly from available training retinal images, it is possible to perform referral diagnostics without detecting individual lesions.
Prior art on automated screening of diabetic retinopathy and direct referral decision shows promising performance; yet most methods build upon complex hand-crafted features whose performance often fails to generalize. OBJECTIVE: We investigate data-driven approaches that extract powerful abstract representations directly from retinal images to provide a reliable referable diabetic retinopathy detector. METHODS: We gradually build the solution based on convolutional neural networks, adding data augmentation, multi-resolution training, robust feature-extraction augmentation, and a patient-basis analysis, testing the effectiveness of each improvement. RESULTS: The proposed method achieved an area under the ROC curve of 98.2% (95% CI: 97.4-98.9%) under a strict cross-dataset protocol designed to test the ability to generalize - training on the Kaggle competition dataset and testing using the Messidor-2 dataset. With a 5 × 2-fold cross-validation protocol, similar results are achieved for Messidor-2 and DR2 datasets, reducing the classification error by over 44% when compared to most published studies in existing literature. CONCLUSION: Additional boost strategies can improve performance substantially, but it is important to evaluate whether the additional (computation- and implementation-) complexity of each improvement is worth its benefits. We also corroborate that novel families of data-driven methods are the state of the art for diabetic retinopathy screening. SIGNIFICANCE: By learning powerful discriminative patterns directly from available training retinal images, it is possible to perform referral diagnostics without detecting individual lesions.