CABNet: Category Attention Block for Imbalanced Diabetic Retinopathy Grading

ADAM Challenge: Detecting Age-Related Macular Degeneration From Fundus Images

Age-related macular degeneration (AMD) is the leading cause of visual impairment among elderly in the world. Early detection of AMD is of great importance, as the vision loss caused by this disease is irreversible and permanent. Color fundus photography is the most cost-effective imaging modality to screen for retinal disorders. Cutting edge deep learning based algorithms have been recently developed for automatically detecting AMD from fundus images. However, there are still lack of a comprehensive annotated dataset and standard evaluation benchmarks. To deal with this issue, we set up the Automatic Detection challenge on Age-related Macular degeneration (ADAM), which was held as a satellite event of the ISBI 2020 conference. The ADAM challenge consisted of four tasks which cover the main aspects of detecting and characterizing AMD from fundus images, including detection of AMD, detection and segmentation of optic disc, localization of fovea, and detection and segmentation of lesions. As part of the ADAM challenge, we have released a comprehensive dataset of 1200 fundus images with AMD diagnostic labels, pixel-wise segmentation masks for both optic disc and AMD-related lesions (drusen, exudates, hemorrhages and scars, among others), as well as the coordinates corresponding to the location of the macular fovea. A uniform evaluation framework has been built to make a fair comparison of different models using this dataset. During the ADAM challenge, 610 results were submitted for online evaluation, with 11 teams finally participating in the onsite challenge. This paper introduces the challenge, the dataset and the evaluation methods, as well as summarizes the participating methods and analyzes their results for each task. In particular, we observed that the ensembling strategy and the incorporation of clinical domain knowledge were the key to improve the performance of the deep learning models.

The Role of Medical Image Modalities and AI in the Early Detection, Diagnosis and Grading of Retinal Diseases: A Survey

Traditional dilated ophthalmoscopy can reveal diseases, such as age-related macular degeneration (AMD), diabetic retinopathy (DR), diabetic macular edema (DME), retinal tear, epiretinal membrane, macular hole, retinal detachment, retinitis pigmentosa, retinal vein occlusion (RVO), and retinal artery occlusion (RAO). Among these diseases, AMD and DR are the major causes of progressive vision loss, while the latter is recognized as a world-wide epidemic. Advances in retinal imaging have improved the diagnosis and management of DR and AMD. In this review article, we focus on the variable imaging modalities for accurate diagnosis, early detection, and staging of both AMD and DR. In addition, the role of artificial intelligence (AI) in providing automated detection, diagnosis, and staging of these diseases will be surveyed. Furthermore, current works are summarized and discussed. Finally, projected future trends are outlined. The work done on this survey indicates the effective role of AI in the early detection, diagnosis, and staging of DR and/or AMD. In the future, more AI solutions will be presented that hold promise for clinical applications.

Weakly-Supervised Learning With Complementary Heatmap for Retinal Disease Detection

There are many types of retinal disease, and accurately detecting these diseases is crucial for proper diagnosis. Convolutional neural networks (CNNs) typically perform well on detection tasks, and the attention module of CNNs can generate heatmaps as visual explanations of the model. However, the generated heatmap can only detect the most discriminative part, which is problematic because many object regions may exist in the region beside the heatmap in an area known as a complementary heatmap. In this study, we developed a method specifically designed multi-retinal diseases detection from fundus images with the complementary heatmap. The proposed CAM-based method is designed for 2D color images of the retina, rather than MRI images or other forms of data. Moreover, unlike other visual images for disease detection, fundus images of multiple retinal diseases have features such as distinguishable lesion region boundaries, overlapped lesion regions between diseases, and specific pathological structures (e.g. scattered blood spots) that lead to mis-classifications. Based on these considerations, we designed two new loss functions, attention-explore loss and attention-refine loss, to generate accurate heatmaps. We select both "bad" and "good" heatmaps based on the prediction score of ground truth and train them with the two loss functions. When the detection accuracy increases, the classification performance of the model is also improved. Experiments on a dataset consisting of five diseases showed that our approach improved both the detection accuracy and the classification accuracy, and the improved heatmaps were closer to the lesion regions than those of current state-of-the-art methods.

Multi-Model Domain Adaptation for Diabetic Retinopathy Classification

Diabetic retinopathy (DR) is one of the most threatening complications in diabetic patients, leading to permanent blindness without timely treatment. However, DR screening is not only a time-consuming task that requires experienced ophthalmologists but also easy to produce misdiagnosis. In recent years, deep learning techniques based on convolutional neural networks have attracted increasing research attention in medical image analysis, especially for DR diagnosis. However, dataset labeling is expensive work and it is necessary for existing deep-learning-based DR detection models. For this study, a novel domain adaptation method (multi-model domain adaptation) is developed for unsupervised DR classification in unlabeled retinal images. At the same time, it only exploits discriminative information from multiple source models without access to any data. In detail, we integrate a weight mechanism into the multi-model-based domain adaptation by measuring the importance of each source domain in a novel way, and a weighted pseudo-labeling strategy is attached to the source feature extractors for training the target DR classification model. Extensive experiments are performed on four source datasets (DDR, IDRiD, Messidor, and Messidor-2) to a target domain APTOS 2019, showing that MMDA produces competitive performance for present state-of-the-art methods for DR classification. As a novel DR detection approach, this article presents a new domain adaptation solution for medical image analysis when the source data is unavailable.

RFormer: Transformer-Based Generative Adversarial Network for Real Fundus Image Restoration on a New Clinical Benchmark

Ophthalmologists have used fundus images to screen and diagnose eye diseases. However, different equipments and ophthalmologists pose large variations to the quality of fundus images. Low-quality (LQ) degraded fundus images easily lead to uncertainty in clinical screening and generally increase the risk of misdiagnosis. Thus, real fundus image restoration is worth studying. Unfortunately, real clinical benchmark has not been explored for this task so far. In this paper, we investigate the real clinical fundus image restoration problem. Firstly, We establish a clinical dataset, Real Fundus (RF), including 120 low- and high-quality (HQ) image pairs. Then we propose a novel Transformer-based Generative Adversarial Network (RFormer) to restore the real degradation of clinical fundus images. The key component in our network is the Window-based Self-Attention Block (WSAB) which captures non-local self-similarity and long-range dependencies. To produce more visually pleasant results, a Transformer-based discriminator is introduced. Extensive experiments on our clinical benchmark show that the proposed RFormer significantly outperforms the state-of-the-art (SOTA) methods. In addition, experiments of downstream tasks such as vessel segmentation and optic disc/cup detection demonstrate that our proposed RFormer benefits clinical fundus image analysis and applications. The dataset, code, and models will be made publicly available at https://github.com/dengzhuo-AI/Real-Fundus.

DeepDRiD: Diabetic Retinopathy-Grading and Image Quality Estimation Challenge

We described a challenge named "Diabetic Retinopathy (DR)-Grading and Image Quality Estimation Challenge" in conjunction with ISBI 2020 to hold three sub-challenges and develop deep learning models for DR image assessment and grading. The scientific community responded positively to the challenge, with 34 submissions from 574 registrations. In the challenge, we provided the DeepDRiD dataset containing 2,000 regular DR images (500 patients) and 256 ultra-widefield images (128 patients), both having DR quality and grading annotations. We discussed details of the top 3 algorithms in each sub-challenges. The weighted kappa for DR grading ranged from 0.93 to 0.82, and the accuracy for image quality evaluation ranged from 0.70 to 0.65. The results showed that image quality assessment can be used as a further target for exploration. We also have released the DeepDRiD dataset on GitHub to help develop automatic systems and improve human judgment in DR screening and diagnosis.

BFENet: A two-stream interaction CNN method for multi-label ophthalmic diseases classification with bilateral fundus images

BACKGROUND AND OBJECTIVE

Early fundus screening and timely treatment of ophthalmology diseases can effectively prevent blindness. Previous studies just focus on fundus images of single eye without utilizing the useful relevant information of the left and right eyes. While clinical ophthalmologists usually use binocular fundus images to help ocular disease diagnosis. Besides, previous works usually target only one ocular diseases at a time. Considering the importance of patient-level bilateral eye diagnosis and multi-label ophthalmic diseases classification, we propose a bilateral feature enhancement network (BFENet) to address the above two problems.

METHODS

We propose a two-stream interactive CNN architecture for multi-label ophthalmic diseases classification with bilateral fundus images. Firstly, we design a feature enhancement module, which makes use of the interaction between bilateral fundus images to strengthen the extracted feature information. Specifically, attention mechanism is used to learn the interdependence between local and global information in the designed interactive architecture for two-stream, which leads to the reweighting of these features, and recover more details. In order to capture more disease characteristics, we further design a novel multiscale module, which enriches the feature maps by superimposing feature information of different resolutions images extracted through dilated convolution.

RESULTS

In the off-site set, the Kappa, F₁, AUC and Final score are 0.535, 0.892, 0.912 and 0.780, respectively. In the on-site set, the Kappa, F₁, AUC and Final score are 0.513, 0.886, 0.903 and 0.767 respectively. Comparing with existing methods, BFENet achieves the best classification performance.

CONCLUSIONS

Comprehensive experiments are conducted to demonstrate the effectiveness of this proposed model. Besides, our method can be extended to similar tasks where the correlation between different images is important.

RTNet: Relation Transformer Network for Diabetic Retinopathy Multi-Lesion Segmentation

Automatic diabetic retinopathy (DR) lesions segmentation makes great sense of assisting ophthalmologists in diagnosis. Although many researches have been conducted on this task, most prior works paid too much attention to the designs of networks instead of considering the pathological association for lesions. Through investigating the pathogenic causes of DR lesions in advance, we found that certain lesions are closed to specific vessels and present relative patterns to each other. Motivated by the observation, we propose a relation transformer block (RTB) to incorporate attention mechanisms at two main levels: a self-attention transformer exploits global dependencies among lesion features, while a cross-attention transformer allows interactions between lesion and vessel features by integrating valuable vascular information to alleviate ambiguity in lesion detection caused by complex fundus structures. In addition, to capture the small lesion patterns first, we propose a global transformer block (GTB) which preserves detailed information in deep network. By integrating the above blocks of dual-branches, our network segments the four kinds of lesions simultaneously. Comprehensive experiments on IDRiD and DDR datasets well demonstrate the superiority of our approach, which achieves competitive performance compared to state-of-the-arts.

The Role of Different Retinal Imaging Modalities in Predicting Progression of Diabetic Retinopathy: A Survey

Diabetic retinopathy (DR) is a devastating condition caused by progressive changes in the retinal microvasculature. It is a leading cause of retinal blindness in people with diabetes. Long periods of uncontrolled blood sugar levels result in endothelial damage, leading to macular edema, altered retinal permeability, retinal ischemia, and neovascularization. In order to facilitate rapid screening and diagnosing, as well as grading of DR, different retinal modalities are utilized. Typically, a computer-aided diagnostic system (CAD) uses retinal images to aid the ophthalmologists in the diagnosis process. These CAD systems use a combination of machine learning (ML) models (e.g., deep learning (DL) approaches) to speed up the diagnosis and grading of DR. In this way, this survey provides a comprehensive overview of different imaging modalities used with ML/DL approaches in the DR diagnosis process. The four imaging modalities that we focused on are fluorescein angiography, fundus photographs, optical coherence tomography (OCT), and OCT angiography (OCTA). In addition, we discuss limitations of the literature that utilizes such modalities for DR diagnosis. In addition, we introduce research gaps and provide suggested solutions for the researchers to resolve. Lastly, we provide a thorough discussion about the challenges and future directions of the current state-of-the-art DL/ML approaches. We also elaborate on how integrating different imaging modalities with the clinical information and demographic data will lead to promising results for the scientists when diagnosing and grading DR. As a result of this article's comparative analysis and discussion, it remains necessary to use DL methods over existing ML models to detect DR in multiple modalities.

Fusing pre-trained convolutional neural networks features for multi-differentiated subtypes of liver cancer on histopathological images

Liver cancer is a malignant tumor with high morbidity and mortality, which has a tremendous negative impact on human survival. However, it is a challenging task to recognize tens of thousands of histopathological images of liver cancer by naked eye, which poses numerous challenges to inexperienced clinicians. In addition, factors such as long time-consuming, tedious work and huge number of images impose a great burden on clinical diagnosis. Therefore, our study combines convolutional neural networks with histopathology images and adopts a feature fusion approach to help clinicians efficiently discriminate the differentiation types of primary hepatocellular carcinoma histopathology images, thus improving their diagnostic efficiency and relieving their work pressure. In this study, for the first time, 73 patients with different differentiation types of primary liver cancer tumors were classified. We performed an adequate classification evaluation of liver cancer differentiation types using four pre-trained deep convolutional neural networks and nine different machine learning (ML) classifiers on a dataset of liver cancer histopathology images with multiple differentiation types. And the test set accuracy, validation set accuracy, running time with different strategies, precision, recall and F1 value were used for adequate comparative evaluation. Proved by experimental results, fusion networks (FuNet) structure is a good choice, which covers both channel attention and spatial attention, and suppresses channel interference with less information. Meanwhile, it can clarify the importance of each spatial location by learning the weights of different locations in space, then apply it to the study of classification of multi-differentiated types of liver cancer. In addition, in most cases, the Stacking-based integrated learning classifier outperforms other ML classifiers in the classification task of multi-differentiation types of liver cancer with the FuNet fusion strategy after dimensionality reduction of the fused features by principle component analysis (PCA) features, and a satisfactory result of 72.46% is achieved in the test set, which has certain practicality.

Classification and Detection of COVID-19 X-Ray Images based on DenseNet and VGG16 Feature Fusion

Since December 2019, the novel coronavirus disease (COVID-19) caused by the syndrome coronavirus 2 (SARS-CoV-2) strain has spread widely around the world and has become a serious global public health problem. For this high-speed infectious disease, the application of X-ray to chest diagnosis plays a key role. In this study, we propose a chest X-ray image classification method based on feature fusion of a dense convolutional network (DenseNet) and a visual geometry group network (VGG16). This paper adds an attention mechanism (global attention machine block and category attention block) to the model to extract deep features. A residual network (ResNet) is used to segment effective image information to quickly achieve accurate classification. The average accuracy of our model in detecting binary classification can reach 98.0%. The average accuracy for three category classification can reach 97.3%. The experimental results show that the proposed model has good results in this work. Therefore, the use of deep learning and feature fusion technology in the classification of chest X-ray images can become an auxiliary tool for clinicians and radiologists.

Multimodal image encoding pre-training for diabetic retinopathy grading

Diabetic retinopathy is an increasingly prevalent eye disorder that can lead to severe vision impairment. The severity grading of the disease using retinal images is key to provide an adequate treatment. However, in order to learn the diverse patterns and complex relations that are required for the grading, deep neural networks require very large annotated datasets that are not always available. This has been typically addressed by reusing networks that were pre-trained for natural image classification, hence relying on additional annotated data from a different domain. In contrast, we propose a novel pre-training approach that takes advantage of unlabeled multimodal visual data commonly available in ophthalmology. The use of multimodal visual data for pre-training purposes has been previously explored by training a network in the prediction of one image modality from another. However, that approach does not ensure a broad understanding of the retinal images, given that the network may exclusively focus on the similarities between modalities while ignoring the differences. Thus, we propose a novel self-supervised pre-training that explicitly teaches the networks to learn the common characteristics between modalities as well as the characteristics that are exclusive to the input modality. This provides a complete comprehension of the input domain and facilitates the training of downstream tasks that require a broad understanding of the retinal images, such as the grading of diabetic retinopathy. To validate and analyze the proposed approach, we performed an exhaustive experimentation on different public datasets. The transfer learning performance for the grading of diabetic retinopathy is evaluated under different settings while also comparing against previous state-of-the-art pre-training approaches. Additionally, a comparison against relevant state-of-the-art works for the detection and grading of diabetic retinopathy is also provided. The results show a satisfactory performance of the proposed approach, which outperforms previous pre-training alternatives in the grading of diabetic retinopathy.

Automatic detection of multiple types of pneumonia: Open dataset and a multi-scale attention network

The quick and precise identification of COVID-19 pneumonia, non-COVID-19 viral pneumonia, bacterial pneumonia, mycoplasma pneumonia, and normal lung on chest CT images play a crucial role in timely quarantine and medical treatment. However, manual identification is subject to potential misinterpretations and time-consumption issues owing the visual similarities of pneumonia lesions. In this study, we propose a novel multi-scale attention network (MSANet) based on a bag of advanced deep learning techniques for the automatic classification of COVID-19 and multiple types of pneumonia. The proposed method can automatically pay attention to discriminative information and multi-scale features of pneumonia lesions for better classification. The experimental results show that the proposed MSANet can achieve an overall precision of 97.31%, recall of 96.18%, F1-score of 96.71%, accuracy of 97.46%, and macro-average area under the receiver operating characteristic curve (AUC) of 0.9981 to distinguish between multiple classes of pneumonia. These promising results indicate that the proposed method can significantly assist physicians and radiologists in medical diagnosis. The dataset is publicly available at https://doi.org/10.17632/rf8x3wp6ss.1.

Optimized Ensemble Machine Learning-Based Diabetic Retinopathy Grading Using Multiple Region of Interest Analysis and Bayesian Approach

Diabetic Retinopathy (DR) is a critical abnormality in the retina mainly caused by diabetes. The early diagnosis of DR is essential to avoid painless blindness. The conventional DR diagnosis is manual and requires skilled Ophthalmologists. The Ophthalmologist’s analyses are subjective to inconsistency and record maintenance issues. Hence, there is a need for other DR diagnosis methods. In this paper, we proposed an AdaBoost algorithm-based ensemble classification approach to classify DR grades. The major objective of the proposed approach is an enhancement of DR classification performance by using optimized features and ensemble machine learning techniques. The proposed method classifies different grades of DR using the Meyer wavelet and retinal vessel-based features extracted from multiple regions of interest of the retina. To improve the predictive accuracy, we used a Bayesian algorithm to optimize the hyper-parameters of the proposed ensemble classifier. The proposed DR grading model was constructed and evaluated by using the MESSIDOR fundus image dataset. In evaluation experiment, the classification outcome of the proposed approach was evaluated by the confusion matrix and receiver operating characteristic (ROC) based metrics. The evaluation experiments show that the proposed approach attained 99.2% precision, 98.2% recall, 99% accuracy, and 0.99 AUC. The experimental findings also indicate that the proposed approach’s classification outcome is significantly better than that of state of art DR classification methods.

Artificial intelligence promotes the diagnosis and screening of diabetic retinopathy

Deep learning evolves into a new form of machine learning technology that is classified under artificial intelligence (AI), which has substantial potential for large-scale healthcare screening and may allow the determination of the most appropriate specific treatment for individual patients. Recent developments in diagnostic technologies facilitated studies on retinal conditions and ocular disease in metabolism and endocrinology. Globally, diabetic retinopathy (DR) is regarded as a major cause of vision loss. Deep learning systems are effective and accurate in the detection of DR from digital fundus photographs or optical coherence tomography. Thus, using AI techniques, systems with high accuracy and efficiency can be developed for diagnosing and screening DR at an early stage and without the resources that are only accessible in special clinics. Deep learning enables early diagnosis with high specificity and sensitivity, which makes decisions based on minimally handcrafted features paving the way for personalized DR progression real-time monitoring and in-time ophthalmic or endocrine therapies. This review will discuss cutting-edge AI algorithms, the automated detecting systems of DR stage grading and feature segmentation, the prediction of DR outcomes and therapeutics, and the ophthalmic indications of other systemic diseases revealed by AI.

Diabetic Retinopathy Diagnosis Based on RA-EfficientNet

The early detection and grade diagnosis of diabetic retinopathy (DR) are very important for the avoidance of blindness, and using deep learning methods to automatically diagnose DR has attracted great attention. However, the small amount of DR data limits its application. To automatically learn the disease’s features and detect DR more accurately, we constructed a DR grade diagnostic model. To realize the model, the authors performed the following steps: firstly, we preprocess the DR images to solve the existing problems in an APTOS 2019 dataset, such as size difference, information redundancy and the data imbalance. Secondly, to extract more valid image features, a new network named RA-EfficientNet is proposed, in which a residual attention (RA) block is added to EfficientNet to extract more features and to solve the problem of small differences between lesions. EfficientNet has been previously trained on the ImageNet dataset, based on transfer learning technology, to overcome the small sample size problem of DR. Lastly, based on the extracted features, two classifiers are designed, one is a 2-grade classifier and the other a 5-grade classifier. The 2-grade classifier can diagnose DR, and the 5-grade classifier provides 5 grades of diagnosis for DR, as follows: 0 for No DR, 1 for mild DR, 2 for moderate, 3 for severe and 4 for proliferative DR. Experiments show that our proposed RA-EfficientNet can achieve better performance, with an accuracy value of 98.36% and a kappa score of 96.72% in a 2-grade classification and an accuracy value of 93.55% and a kappa score of 91.93% in a 5-grade classification. The results indicate that the proposed model effectively improves DR detection efficiency and resolves the existing limitation of manual feature extraction.

C+EffxNet: A novel hybrid approach for COVID-19 diagnosis on CT images based on CBAM and EfficientNet

COVID-19, one of the biggest diseases of our age, continues to spread rapidly around the world. Studies continue rapidly for the diagnosis and treatment of this disease. It is of great importance that individuals who are infected with this virus be isolated from the rest of the society so that the disease does not spread further. In addition to the tests performed in the detection process of the patients, X-ray and computed tomography are also used. In this study, a new hybrid model that can diagnose COVID-19 from computed tomography images created using EfficientNet, one of the current deep learning models, with a model consisting of attention blocks is proposed. In the first step of this new model, channel attention, spatial attention, and residual blocks are used to extract the most important features from the images. The extracted features are combined in accordance with the hyper-column technique. The combined features are given as input to the EfficientNet models in the second step of the model. The deep features obtained from this proposed hybrid model were classified with the Support Vector Machine classifier after feature selection. Principal Components Analysis was used for feature selection. The approach can accurately predict COVID-19 with a 99% accuracy rate. The first four versions of EfficientNet are used in the approach. In addition, Bayesian optimization was used in the hyper parameter estimation of the Support Vector Machine classifier. Comparative performance analysis of the approach with other approaches in the field is given.

Automated Detection and Diagnosis of Diabetic Retinopathy: A Comprehensive Survey

Diabetic Retinopathy (DR) is a leading cause of vision loss in the world. In the past few years, artificial intelligence (AI) based approaches have been used to detect and grade DR. Early detection enables appropriate treatment and thus prevents vision loss. For this purpose, both fundus and optical coherence tomography (OCT) images are used to image the retina. Next, Deep-learning (DL)-/machine-learning (ML)-based approaches make it possible to extract features from the images and to detect the presence of DR, grade its severity and segment associated lesions. This review covers the literature dealing with AI approaches to DR such as ML and DL in classification and segmentation that have been published in the open literature within six years (2016-2021). In addition, a comprehensive list of available DR datasets is reported. This list was constructed using both the PICO (P-Patient, I-Intervention, C-Control, O-Outcome) and Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) 2009 search strategies. We summarize a total of 114 published articles which conformed to the scope of the review. In addition, a list of 43 major datasets is presented.

Multi-level semantic adaptation for few-shot segmentation on cardiac image sequences

Obtaining manual labels is time-consuming and labor-intensive on cardiac image sequences. Few-shot segmentation can utilize limited labels to learn new tasks. However, it suffers from two challenges: spatial-temporal distribution bias and long-term information bias. These challenges derive from the impact of the time dimension on cardiac image sequences, resulting in serious over-adaptation. In this paper, we propose the multi-level semantic adaptation (MSA) for few-shot segmentation on cardiac image sequences. The MSA addresses the two biases by exploring the domain adaptation and the weight adaptation on the semantic features in multiple levels, including sequence-level, frame-level, and pixel-level. First, the MSA proposes the dual-level feature adjustment for domain adaptation in spatial and temporal directions. This adjustment explicitly aligns the frame-level feature and the sequence-level feature to improve the model adaptation on diverse modalities. Second, the MSA explores the hierarchical attention metric for weight adaptation in the frame-level feature and the pixel-level feature. This metric focuses on the similar frame and the target region to promote the model discrimination on the border features. The extensive experiments demonstrate that our MSA is effective in few-shot segmentation on cardiac image sequences with three modalities, i.e. MR, CT, and Echo (e.g. the average Dice is 0.9243), as well as superior to the ten state-of-the-art methods.

Harmony Loss for Unbalanced Prediction

In medical image analysis, in order to reduce the impact of unbalanced data sets on data-driven deep learning models, according to the characteristic that the area under the Precision-Recall curve (AUC_PR) is sensitive to each category of samples, a novel Harmony loss function with fast convergence speed and high stability was constructed. Since AUC_PR needs to be calculated in discrete domain, in order to ensure the continuous differentiability and gradient existence of the Harmony loss, first, the Logistic function was used to approximate the Logical function in AUC_PR. Then, to improve the optimization speed of the Harmony loss during model training, a method of manually setting a certain number of classification thresholds was proposed to further approximate the calculation of AUC_PR. After the above two approximate calculation processes, the Harmony loss with stable gradient and high computational efficiency was designed. In the optimization process of the model, since Harmony loss can reconcile recall and precision of each category under different classification thresholds, thereby, it can not only improve the models ability to recognize categories with less samples, but also maintain the stability of the training curve. To comprehensively evaluate the effects of Harmony loss function, we performed experiments on image 3D reconstruction, 2D segmentation, and unbalanced classification tasks. Experimental results showed that the Harmony loss achieved the state-of-the-art results on four unbalanced data sets. Moreover, the Harmony loss can be easily combined with existing loss functions, and is suitable for most common deep learning models.

Diabetic Retinopathy Fundus Image Classification and Lesions Localization System Using Deep Learning

Diabetic retinopathy (DR) is a disease resulting from diabetes complications, causing non-reversible damage to retina blood vessels. DR is a leading cause of blindness if not detected early. The currently available DR treatments are limited to stopping or delaying the deterioration of sight, highlighting the importance of regular scanning using high-efficiency computer-based systems to diagnose cases early. The current work presented fully automatic diagnosis systems that exceed manual techniques to avoid misdiagnosis, reducing time, effort and cost. The proposed system classifies DR images into five stages—no-DR, mild, moderate, severe and proliferative DR—as well as localizing the affected lesions on retain surface. The system comprises two deep learning-based models. The first model (CNN512) used the whole image as an input to the CNN model to classify it into one of the five DR stages. It achieved an accuracy of 88.6% and 84.1% on the DDR and the APTOS Kaggle 2019 public datasets, respectively, compared to the state-of-the-art results. Simultaneously, the second model used an adopted YOLOv3 model to detect and localize the DR lesions, achieving a 0.216 mAP in lesion localization on the DDR dataset, which improves the current state-of-the-art results. Finally, both of the proposed structures, CNN512 and YOLOv3, were fused to classify DR images and localize DR lesions, obtaining an accuracy of 89% with 89% sensitivity, 97.3 specificity and that exceeds the current state-of-the-art results.

Cross-attention multi-branch network for fundus diseases classification using SLO images

Fundus diseases classification is vital for the health of human beings. However, most of existing methods detect diseases by means of single angle fundus images, which lead to the lack of pathological information. To address this limitation, this paper proposes a novel deep learning method to complete different fundus diseases classification tasks using ultra-wide field scanning laser ophthalmoscopy (SLO) images, which have an ultra-wide field view of 180-200˚. The proposed deep model consists of multi-branch network, atrous spatial pyramid pooling module (ASPP), cross-attention and depth-wise attention module. Specifically, the multi-branch network employs the ResNet-34 model as the backbone to extract feature information, where the ResNet-34 model with two-branch is followed by the ASPP module to extract multi-scale spatial contextual features by setting different dilated rates. The depth-wise attention module can provide the global attention map from the multi-branch network, which enables the network to focus on the salient targets of interest. The cross-attention module adopts the cross-fusion mode to fuse the channel and spatial attention maps from the ResNet-34 model with two-branch, which can enhance the representation ability of the disease-specific features. The extensive experiments on our collected SLO images and two publicly available datasets demonstrate that the proposed method can outperform the state-of-the-art methods and achieve quite promising classification performance of the fundus diseases.