Sharif Amit Kamran, Sourajit Saha, Ali Shihab Sabbir, Alireza Tavakkoli
Diagnosing different retinal diseases from Spectral Domain Optical Coherence Tomography (SD-OCT) images is a challenging task. Different automated approaches such as image processing, machine learning and deep learning algorithms have been used for early detection and diagnosis of retinal diseases. Unfortunately, these are prone to error and computational inefficiency, which requires further intervention from human experts. In this paper, we propose a novel convolution neural network architecture to successfully distinguish between different degeneration of retinal layers and their underlying causes. The proposed novel architecture outperforms other classification models while addressing the issue of gradient explosion. Our approach reaches near perfect accuracy of 99.8% and 100% for two separately available Retinal SD-OCT data-set respectively. Additionally, our architecture predicts retinal diseases in real time while outperforming human diagnosticians.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Disease Prediction | OCT2017 | Acc | 99.8 | OpticNet-71 |
| Disease Prediction | OCT2017 | Sensitivity | 99.8 | OpticNet-71 |
| Disease Prediction | OCT2017 | Acc | 99.3 | ResNet50-v1 |
| Disease Prediction | OCT2017 | Sensitivity | 99.3 | ResNet50-v1 |
| Disease Prediction | OCT2017 | Acc | 99.3 | ResNet50-v1 |
| Disease Prediction | OCT2017 | Sensitivity | 99.3 | ResNet50-v1 |
| Disease Prediction | OCT2017 | Acc | 96.6 | InceptionV3 |
| Disease Prediction | OCT2017 | Sensitivity | 97.8 | InceptionV3 |
| Disease Prediction | OCT2017 | Acc | 96.6 | InceptionV3 |
| Disease Prediction | OCT2017 | Sensitivity | 97.8 | InceptionV3 |
| Disease Prediction | OCT2017 | Acc | 93.4 | InceptionV3 (limited) |
| Disease Prediction | OCT2017 | Sensitivity | 96.6 | InceptionV3 (limited) |
| Disease Prediction | OCT2017 | Acc | 93.4 | InceptionV3 (limited) |
| Disease Prediction | OCT2017 | Sensitivity | 96.6 | InceptionV3 (limited) |
| Disease Prediction | Srinivasan2014 | Acc | 100 | OpticNet-71 |
| Disease Prediction | Srinivasan2014 | Acc | 99.36 | Xception |
| Disease Prediction | Srinivasan2014 | Acc | 99.36 | Xception |
| Disease Prediction | Srinivasan2014 | Acc | 97.46 | MobileNet-v2 |
| Disease Prediction | Srinivasan2014 | Acc | 97.46 | MobileNet-v2 |
| Disease Prediction | Srinivasan2014 | Acc | 96 | Karri et al. |
| Disease Prediction | Srinivasan2014 | Acc | 94.92 | ResNet50-v1 |
| Disease Prediction | Srinivasan2014 | Acc | 94.92 | ResNet50-v1 |
| Disease Prediction | Srinivasan2014 | Acc | 93 | Awais et al. |
| Disease Prediction | Srinivasan2014 | Acc | 87.63 | Lee et al. |
| Disease Prediction | Srinivasan2014 | Acc | 87.63 | Lee et al. |
| Medical Diagnosis | OCT2017 | Acc | 99.8 | OpticNet-71 |
| Medical Diagnosis | OCT2017 | Sensitivity | 99.8 | OpticNet-71 |
| Medical Diagnosis | OCT2017 | Acc | 99.3 | ResNet50-v1 |
| Medical Diagnosis | OCT2017 | Sensitivity | 99.3 | ResNet50-v1 |
| Medical Diagnosis | OCT2017 | Acc | 99.3 | ResNet50-v1 |
| Medical Diagnosis | OCT2017 | Sensitivity | 99.3 | ResNet50-v1 |
| Medical Diagnosis | OCT2017 | Acc | 96.6 | InceptionV3 |
| Medical Diagnosis | OCT2017 | Sensitivity | 97.8 | InceptionV3 |
| Medical Diagnosis | OCT2017 | Acc | 96.6 | InceptionV3 |
| Medical Diagnosis | OCT2017 | Sensitivity | 97.8 | InceptionV3 |
| Medical Diagnosis | OCT2017 | Acc | 93.4 | InceptionV3 (limited) |
| Medical Diagnosis | OCT2017 | Sensitivity | 96.6 | InceptionV3 (limited) |
| Medical Diagnosis | OCT2017 | Acc | 93.4 | InceptionV3 (limited) |
| Medical Diagnosis | OCT2017 | Sensitivity | 96.6 | InceptionV3 (limited) |
| Medical Diagnosis | Srinivasan2014 | Acc | 100 | OpticNet-71 |
| Medical Diagnosis | Srinivasan2014 | Acc | 99.36 | Xception |
| Medical Diagnosis | Srinivasan2014 | Acc | 99.36 | Xception |
| Medical Diagnosis | Srinivasan2014 | Acc | 97.46 | MobileNet-v2 |
| Medical Diagnosis | Srinivasan2014 | Acc | 97.46 | MobileNet-v2 |
| Medical Diagnosis | Srinivasan2014 | Acc | 96 | Karri et al. |
| Medical Diagnosis | Srinivasan2014 | Acc | 94.92 | ResNet50-v1 |
| Medical Diagnosis | Srinivasan2014 | Acc | 94.92 | ResNet50-v1 |
| Medical Diagnosis | Srinivasan2014 | Acc | 93 | Awais et al. |
| Medical Diagnosis | Srinivasan2014 | Acc | 87.63 | Lee et al. |
| Medical Diagnosis | Srinivasan2014 | Acc | 87.63 | Lee et al. |