Attentional Mechanisms and Improved Residual Networks for Diabetic Retinopathy Severity Classification

A Self-Supervised Equivariant Refinement Classification Network for Diabetic Retinopathy Classification

Diabetic retinopathy (DR) is a retinal disease caused by diabetes. If there is no intervention, it may even lead to blindness. Therefore, the detection of diabetic retinopathy is of great significance for preventing blindness in patients. Most of the existing DR detection methods use supervised methods, which usually require a large number of accurate pixel-level annotations. To solve this problem, we propose a self-supervised Equivariant Refinement Classification Network (ERCN) for DR classification. First, we use an unsupervised contrast pre-training network to learn a more generalized representation. Secondly, the class activation map (CAM) is refined by self-supervision learning. It first uses a spatial masking method to suppress low-confidence predictions, and then uses the feature similarity between pixels to encourage fine-grained activation to achieve more accurate positioning of the lesion. We propose a hybrid equivariant regularization loss to alleviate the degradation caused by the local minimum in the CAM refinement process. To further improve the classification accuracy, we propose an attention-based multi-instance learning (MIL), which weights each element of the feature map as an instance, which is more effective than the traditional patch-based instance extraction method. We evaluate our method on the EyePACS and DAVIS datasets and achieved 87.4% test accuracy in the EyePACS dataset and 88.7% test accuracy in the DAVIS dataset. It shows that the proposed method achieves better performance in DR detection compared with other state-of-the-art methods in self-supervised DR detection.

A graph convolutional network with dynamic weight fusion of multi-scale local features for diabetic retinopathy grading

Diabetic retinopathy (DR) is a serious ocular complication that can pose a serious risk to a patient's vision and overall health. Currently, the automatic grading of DR is mainly using deep learning techniques. However, the lesion information in DR images is complex, variable in shape and size, and randomly distributed in the images, which leads to some shortcomings of the current research methods, i.e., it is difficult to effectively extract the information of these various features, and it is difficult to establish the connection between the lesion information in different regions. To address these shortcomings, we design a multi-scale dynamic fusion (MSDF) module and combine it with graph convolution operations to propose a multi-scale dynamic graph convolutional network (MDGNet) in this paper. MDGNet firstly uses convolution kernels with different sizes to extract features with different shapes and sizes in the lesion regions, and then automatically learns the corresponding weights for feature fusion according to the contribution of different features to model grading. Finally, the graph convolution operation is used to link the lesion features in different regions. As a result, our proposed method can effectively combine local and global features, which is beneficial for the correct DR grading. We evaluate the effectiveness of method on two publicly available datasets, namely APTOS and DDR. Extensive experiments demonstrate that our proposed MDGNet achieves the best grading results on APTOS and DDR, and is more accurate and diverse for the extraction of lesion information.

IMoVR-Net: A robust interpretable network for multi-ocular lesion recognition from TAO facial images

Thyroid-associated ophthalmopathy (TAO) is an organ-specific autoimmune disease that seriously affects patient's life and health. However, the early diagnosis of TAO is highly dependent on the physician's subjective experience. Moreover, the currently proposed deep learning networks for eye diseases do not provide robust interpretability concerning feature learning paradigm, model structure, and the number of neurons. But the mentioned components are very important for model interpretability and are key factors that severely affect the transparency of the model. Therefore, a robust interpretable multi-orientation visual recognition network (IMoVR-Net) for TAO multi-ocular lesion recognition is proposed in this paper. Firstly, a multi-orientation visual cascaded encoder composed of the DenseGabor module and the dilated Gabor convolution group is proposed to achieve the fine extraction of multi-directional TAO lesion features by using a novel feature learning paradigm called alternating filtering. Besides, combining information theory and topology tool, an optimization rule based on topological energy entropy is proposed to provide a solid interpretable theory for determining the model structure. Finally, a clustering correlation analysis method is developed to accomplish the determination of the number of convolutional hidden neurons, providing robust interpretability for the selection of the number of neurons. Compared to other advanced models, the IMoVR-Net achieved state-of-the-art performance on different TAO ocular datasets with an average accuracy, sensitivity, precision, and F1 score of 0.878, 87.27 %, 0.875, and 87.39 %, respectively. The IMoVR-Net has good clinical application prospects due to its strong recognition ability and robust interpretability in feature extraction paradigm, model structure, and number of convolutional neurons.