Weakly-supervised localization of diabetic retinopathy lesions in retinal fundus images

Gondal, Waleed M.; Köhler, Jan M.; Grzeszick, René; Fink, Gernot A.; Hirsch, Michael

doi:10.48550/arxiv.1706.09634

Cited by 3 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Examples include the work of Feng et al [12] on pulmonary nodule localisation in CT, the work of Ge et al [18] on skin disease recognition. Other examples are [69], [19]. It is important to note that CAM is restricted in the resolution of its visual attributions by the resolution of the last feature map.…”

Section: Visual Attributionmentioning

confidence: 99%

“…In this paper we address the problem of visual attribution, which we define as detecting and visualising evidence of a particular category in an image. Pinpointing all evidence of a class is important for a variety of tasks such as weakly supervised localisation or segmentation of structures [43,45,67], and better understanding disease effects, and physiological or pathological processes in medical images [69,18,12,19,13,28,56,31,32,65].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Visual Feature Attribution Using Wasserstein GANs

Baumgartner

Koch

Tezcan

et al. 2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

128

188

View full text Add to dashboard Cite

Attributing the pixels of an input image to a certain category is an important and well-studied problem in computer vision, with applications ranging from weakly supervised localisation to understanding hidden effects in the data. In recent years, approaches based on interpreting a previously trained neural network classifier have become the de facto state-of-the-art and are commonly used on medical as well as natural image datasets. In this paper, we discuss a limitation of these approaches which may lead to only a subset of the category specific features being detected. To address this problem we develop a novel feature attribution technique based on Wasserstein Generative Adversarial Networks (WGAN), which does not suffer from this limitation. We show that our proposed method performs substantially better than the state-of-the-art for visual attribution on a synthetic dataset and on real 3D neuroimaging data from patients with mild cognitive impairment (MCI) and Alzheimer's disease (AD). For AD patients the method produces compellingly realistic disease effect maps which are very close to the observed effects.

show abstract

Section: Visual Attributionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Visual Feature Attribution Using Wasserstein GANs

Baumgartner

Koch

Tezcan

et al. 2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

128

188

View full text Add to dashboard Cite

show abstract

“…The total kappa score accuracy for analysis is 0.74, with 10% of the photos utilized for validation. A CNN model has been proposed by Gondal et al (2017) for the detection of referable DR. Based on binary classi cation, the CNN model performs with a sensitivity of 93.6% and speci city of 97.6% on DiaretDB1. Pratt et al (2016) presented a CNN structure for identifying ve phases but could not effectively categorize the moderate stage owing to the complexity of the design.…”

Section: Literature Reviewmentioning

confidence: 99%

Automatic diabetic retinopathy detection using an ensemble learning approach and classifiers with self-adjusting weights

Nadda,

Singh,

Shrivastava

2024

Preprint

View full text Add to dashboard Cite

Diabetic retinopathy (DR), a common eye disease, can cause extreme damage to the vitals of the patients or even blindness. This disease can cause vision impairment and possibly full blindness in diabetic patients who do not receive the correct diagnosis and treatment in the early stages. Diabetic retinopathy must be detected early since the disease will gradually damage the eye. A computer-aided prognosis-based technique is currently being used to assist clinicians in identifying DR in its earliest stages. The existing approaches for classifying stages of DR have an inadequate ability to reliably detect early stages due to their inability to capture the complex underlying features. The present study aimed to automatically identify the stages of DR disease by evaluating retina image samples using a proposed ensemble machine learning approach with self-adjusting classifier weights. To improve the categorization of various stages of DR, the present study trained an ensemble of three deep Convolutional Neural Network (CNN) models and two machine learning models (ResNet50, Densenet121, Squeezenet1_0, SVM, and decision tree as base learners) using the openly accessible Kaggle dataset of retina pictures, to encode rich characteristics. The study results demonstrate that, in contrast to existing approaches, the proposed model identifies all stages of DR and outperforms state-of-the-art methods on the identical Kaggle dataset, with an accuracy of 98.35%.

show abstract

“…Recently, Gondal et.al. [2] leveraged CAM maps to highlight the lesion areas for diabetic retinopathy; however, this approach provides no interpretation of grading information and thus is limited in providing clinical insight on grading decisions. Motivated by the need for clinical interpretability, we propose CLEAR-DR, a novel interpretable CAD system based on the notion of CLass-Enhanced Attentive Response Discovery Radiomics for the purpose of clinical decision support for diabetic retinopathy.…”

Section: Introductionmentioning

confidence: 99%

Discovery Radiomics With CLEAR-DR: Interpretable Computer Aided Diagnosis of Diabetic Retinopathy

2019

View full text Add to dashboard Cite

Objective: Radiomics-driven Computer Aided Diagnosis (CAD) has shown considerable promise in recent years as a potential tool for improving clinical decision support in medical oncology, particularly those based around the concept of Discovery Radiomics, where radiomic sequencers are discovered through the analysis of medical imaging data. One of the main limitations with current CAD approaches is that it is very difficult to gain insight or rationale as to how decisions are made, thus limiting their utility to clinicians. Methods: In this study, we propose CLEAR-DR, a novel interpretable CAD system based on the notion of CLass-Enhanced Attentive Response Discovery Radiomics for the purpose of clinical decision support for diabetic retinopathy. Results: In addition to disease grading via the discovered deep radiomic sequencer, the CLEAR-DR system also produces a visual interpretation of the decision-making process to provide better insight and understanding into the decisionmaking process of the system. Conclusion: We demonstrate the effectiveness and utility of the proposed CLEAR-DR system of enhancing the interpretability of diagnostic grading results for the application of diabetic retinopathy grading. Significance: CLEAR-DR can act as a potential powerful tool to address the uninterpretability issue of current CAD systems, thus improving their utility to clinicians.

show abstract

Weakly-supervised localization of diabetic retinopathy lesions in retinal fundus images

Cited by 3 publications

References 0 publications

Visual Feature Attribution Using Wasserstein GANs

Visual Feature Attribution Using Wasserstein GANs

Automatic diabetic retinopathy detection using an ensemble learning approach and classifiers with self-adjusting weights

Discovery Radiomics With CLEAR-DR: Interpretable Computer Aided Diagnosis of Diabetic Retinopathy

Contact Info

Product

Resources

About