Deep Bayesian baseline for segmenting diabetic retinopathy lesions: Advances and challenges

Improved diabetic retinopathy severity classification using squeeze-and-excitation and sparse light weight multi-level attention u-net with transfer learning from xception

AIMS

Diabetic Retinopathy (DR) is a significant cause of vision loss in diabetic patients, making early detection and accurate severity classification essential for effective management and prevention. This study aims to develop an enhanced DR severity classification approach using advanced model architectures and transfer learning to improve diagnostic accuracy and support better patient care.

METHODS

We propose a novel model, Xception Squeeze-and-Excitation Sparse Lightweight Multi-Level Attention U-Net (XceSE_SparseLwMLA-UNet), designed to classify DR severity using fundus images from the Messidor 1 and Messidor 2 datasets. The XceSE_SparseLwMLA-UNet integrates several advanced mechanisms: the Squeeze-and-Excitation (SE) mechanism for adaptive feature recalibration, the Sparse Lightweight Multi-Level Attention (SparseLwMLA) mechanism for effective contextual information integration, and transfer learning from the Xception architecture to enhance feature extraction capabilities. The SE mechanism refines channel-wise feature responses, while SparseLwMLA enhances the model's ability to identify complex DR patterns. Transfer learning utilizes pre-trained weights from Xception to improve generalization across DR severity levels.

RESULTS

The proposed XceSE_SparseLwMLA-UNet model demonstrates superior performance in DR severity classification, achieving higher accuracy and improved multi-class F1 scores compared to existing models. The model's color-coded segmentation outputs offer interpretable visual representations, aiding medical professionals in assessing DR severity levels.

CONCLUSIONS

The XceSE_SparseLwMLA-UNet model shows promise for advancing early DR diagnosis and management by enhancing classification accuracy and providing valuable visual insights. Its integration of advanced architectural features and transfer learning contributes to better patient care and improved visual health outcomes.

Lightweight and multi-lesion segmentation model for diabetic retinopathy based on the fusion of mixed attention and ghost feature mapping

Diabetic retinopathy is the main cause of blindness, and lesion segmentation is an important basic work for the diagnosis of this disease. The main lesions include soft and hard exudates, microaneurysms, and hemorrhages. However, the segmentation of these four types of lesions is difficult because of their variability in size and contrast, and high intertype similarity. Currently, many network models have problems, such as a large number of parameters and complex calculations, and most segmentation models for diabetic retinopathy focus only on one type of lesion. In this study, a lightweight algorithm based on BiSeNet V2 was proposed for the segmentation of multiple lesions in diabetic retinopathy fundus. First, a hybrid attention module was embedded in the semantic branch of BiSeNet V2 for 8- and 16-fold downsampling, which helped reassign deep feature-map weights and enhanced the ability to extract local key features. Second, a ghost feature-mapping unit was used to optimize the traditional convolution layers and further reduce the computational cost. Third, a new loss function based on the dynamic threshold loss function was applied to supervise the training by adjusting the training weights of the high-loss difficult samples, which enhanced the model's attention to small goals. In experiments on the IDRiD dataset, we conducted an ablation study to verify the effectiveness of each component and compared the proposed model, BiSeNet V2-Pro, with several state-of-the-art models. In comparison with the baseline BiSeNet V2, the segmentation performance of BiSeNet V2-Pro improved by 12.17 %, 11.44 %, and 8.49 % in terms of Sensitivity (SEN), Intersection over Union (IoU), and Dice coefficient (DICE), respectively. Specifically, IoU of MA reaches 0.5716. Compared with other methods, the segmentation speed was significantly improved while ensuring segmentation accuracy, and the number of model parameters was lower. These results demonstrate the superiority of BiSeNet V2-Pro in the multi-lesion segmentation of diabetic retinopathy.

Level-set based adaptive-active contour segmentation technique with long short-term memory for diabetic retinopathy classification

Diabetic Retinopathy (DR) is a major type of eye defect that is caused by abnormalities in the blood vessels within the retinal tissue. Early detection by automatic approach using modern methodologies helps prevent consequences like vision loss. So, this research has developed an effective segmentation approach known as Level-set Based Adaptive-active Contour Segmentation (LBACS) to segment the images by improving the boundary conditions and detecting the edges using Level Set Method with Improved Boundary Indicator Function (LSMIBIF) and Adaptive-Active Counter Model (AACM). For evaluating the DR system, the information is collected from the publically available datasets named as Indian Diabetic Retinopathy Image Dataset (IDRiD) and Diabetic Retinopathy Database 1 (DIARETDB 1). Then the collected images are pre-processed using a Gaussian filter, edge detection sharpening, Contrast enhancement, and Luminosity enhancement to eliminate the noises/interferences, and data imbalance that exists in the available dataset. After that, the noise-free data are processed for segmentation by using the Level set-based active contour segmentation technique. Then, the segmented images are given to the feature extraction stage where Gray Level Co-occurrence Matrix (GLCM), Local ternary, and binary patterns are employed to extract the features from the segmented image. Finally, extracted features are given as input to the classification stage where Long Short-Term Memory (LSTM) is utilized to categorize various classes of DR. The result analysis evidently shows that the proposed LBACS-LSTM achieved better results in overall metrics. The accuracy of the proposed LBACS-LSTM for IDRiD and DIARETDB 1 datasets is 99.43% and 97.39%, respectively which is comparably higher than the existing approaches such as Three-dimensional semantic model, Delimiting Segmentation Approach Using Knowledge Learning (DSA-KL), K-Nearest Neighbor (KNN), Computer aided method and Chronological Tunicate Swarm Algorithm with Stacked Auto Encoder (CTSA-SAE).

A generative adversarial neural network with multi-attention feature extraction for fundus lesion segmentation

PURPOSE

Fundus lesion segmentation determines the location and size of diabetes retinopathy in fundus image, which assists doctors in developing the best eye treatment plan. However, owing to the scattered distribution and the similarity of lesions, it is extremely difficult to extract representative lesions feature and accurately segment lesions area.

METHODS

To solve the thorny problem, a generative adversarial network with multi-attention feature extraction is developed to segment diabetic retinopathy region. The main contributions are as follows: (1) An improved residual U-Net network combining with self-attention mechanism is designed as generative network to fully extract local and global feature of lesions while reducing the loss of key feature information. Considering the correlation between the same lesions feature of different samples, external attention mechanism is introduced in the residual U-Net network to focus on the relevant features of the same lesions in different samples throughout the entire dataset. (2) A discriminative network based on the PatchGAN structure is designed to further enhance the segmentation ability of generation network by discriminating between true and false samples.

RESULTS

The proposed network is evaluated on the public dataset IDRiD, which achieved the Dice correlation coefficients of 75.7%, 76.53%, 50.06%, and 45.89% for EX, SE, MA, and HE, respectively.

CONCLUSION

The experimental results show the generative adversarial neural network qualified for accurate segmentation of diabetic retinopathy from fundus image well.

A hybrid approach for diagnosing diabetic retinopathy from fundus image exploiting deep features

One of the major causes of blindness in human beings is the diabetic retinopathy (DR). To prevent blindness, early detection of DR is therefore necessary. In this paper, a hybrid model is proposed for diagnosing DR from fundus images. A combination of morphological image processing and Inception v3 deep learning techniques are exploited to detect DR as well as to classify healthy, mild non-proliferative DR (NPDR), moderate NPDR, severe NPDR, and proliferative DR (PDR). The proposed algorithm was carried out in several steps such as segmentation of blood vessels, localization and removal of optic disc, and macula, abnormal features detection (microaneurysms, hemorrhages, and neovascularization), and classification. Microaneurysms and hemorrhages that appear in the retina are the early signs of DR. In this work, we have detected microaneurysms and hemorrhages by applying dynamic contrast limited adaptive histogram equalization and threshold value on overlapping patched images. An overall accuracy of 96.83% is obtained to classify DR into five different stages. The better performance demonstrates the effectiveness and novelty of the proposed work as compared to the recent reported work.

iDRNA-ITF: identifying DNA- and RNA-binding residues in proteins based on induction and transfer framework

Protein-DNA and protein-RNA interactions are involved in many biological activities. In the post-genome era, accurate identification of DNA- and RNA-binding residues in protein sequences is of great significance for studying protein functions and promoting new drug design and development. Therefore, some sequence-based computational methods have been proposed for identifying DNA- and RNA-binding residues. However, they failed to fully utilize the functional properties of residues, leading to limited prediction performance. In this paper, a sequence-based method iDRNA-ITF was proposed to incorporate the functional properties in residue representation by using an induction and transfer framework. The properties of nucleic acid-binding residues were induced by the nucleic acid-binding residue feature extraction network, and then transferred into the feature integration modules of the DNA-binding residue prediction network and the RNA-binding residue prediction network for the final prediction. Experimental results on four test sets demonstrate that iDRNA-ITF achieves the state-of-the-art performance, outperforming the other existing sequence-based methods. The webserver of iDRNA-ITF is freely available at http://bliulab.net/iDRNA-ITF.

Collaborative learning of weakly-supervised domain adaptation for diabetic retinopathy grading on retinal images

Early detection and treatment of diabetic retinopathy (DR) can significantly reduce the risk of vision loss in patients. In essence, we are faced with two challenges: (i) how to simultaneously achieve domain adaptation from the different domains and (ii) how to build an interpretable multi-instance learning (MIL) on the target domain in an end-to-end framework. In this paper, we address these issues and propose a unified weakly-supervised domain adaptation framework, which consists of three components: domain adaptation, instance progressive discriminator and multi-instance learning with attention. The method models the relationship between the patches and images in the target domain with a multi-instance learning scheme and an attention mechanism. Meanwhile, it incorporates all available information from both source and target domains for a jointly learning strategy. We validate the performance of the proposed framework for DR grading on the Messidor dataset and the large-scale Eyepacs dataset. The experimental results demonstrate that it achieves an average accuracy of 0.949 (95% CI 0.931-0.958)/0.764 (95% CI 0.755-0.772) and an average AUC value of 0.958 (95% CI 0.945-0.962)/0.749 (95% CI 0.732-0.761) for binary-class/multi-class classification tasks on the Messidor dataset. Moreover, the proposed method achieves an accuracy of 0.887 and a quadratic weighted kappa score value of 0.860 on the Eyepacs dataset, outperforming the state-of-the-art approaches. Comprehensive experiments confirm the effectiveness of the approach in terms of both grading performance and interpretability. The source code is available at https://github.com/HouQingshan/WAD-Net.