Data fusion for multi-lesion Diabetic Retinopathy detection

Jelinek, Herbert F.; Pires, Ramon; Goldenstein, Siome; Wainer, Jacques; Bossomaier, Terry; Rocha, Armando Freitas da

doi:10.1109/cbms.2012.6266342

Cited by 18 publications

(25 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From all the images, 595 were classified as containing no sign of DR and 482 as showing pathological signs. In this case, we are only aware of a work addressing the task of DR detection on this dataset [45]. Since DR1 contains ground-truth regarding the presence of different lesions within each pathological image, DR detection is achieved by training separate detectors for each of them and then fusing the results with a meta-classifier.…”

Section: A Dr Detection Performance Evaluationmentioning

confidence: 99%

“…Since DR1 contains ground-truth regarding the presence of different lesions within each pathological image, DR detection is achieved by training separate detectors for each of them and then fusing the results with a meta-classifier. Performance comparison with the results in [45] is shown in Table 3. In this case, the proposed technique clearly outperforms the method proposed in [45], which confirms the generality of our approach.…”

Section: A Dr Detection Performance Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

A Weakly-Supervised Framework for Interpretable Diabetic Retinopathy Detection on Retinal Images

et al. 2018

View full text Add to dashboard Cite

Diabetic retinopathy (DR) detection is a critical retinal image analysis task in the context of early blindness prevention. Unfortunately, in order to train a model to accurately detect DR based on the presence of different retinal lesions, typically a dataset with medical expert's annotations at the pixel level is needed. In this paper, a new methodology based on the multiple instance learning (MIL) framework is developed in order to overcome this necessity by leveraging the implicit information present on annotations made at the image level. Contrary to previous MIL-based DR detection systems, the main contribution of the proposed technique is the joint optimization of the instance encoding and the image classification stages. In this way, more useful mid-level representations of pathological images can be obtained. The explainability of the model decisions is further enhanced by means of a new loss function enforcing appropriate instance and mid-level representations. The proposed technique achieves comparable or better results than other recently proposed methods, with 90% area under the receiver operating characteristic curve (AUC) on Messidor, 93% AUC on DR1, and 96% AUC on DR2, while improving the interpretability of the produced decisions.INDEX TERMS Multiple instance learning, diabetic retinopathy detection, bag of visual words, retinal image analysis.

show abstract

Section: A Dr Detection Performance Evaluationmentioning

confidence: 99%

Section: A Dr Detection Performance Evaluationmentioning

confidence: 99%

A Weakly-Supervised Framework for Interpretable Diabetic Retinopathy Detection on Retinal Images

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Jelinek et al [31] employed data fusion for the detection of multiple lesions in the retinal images. The authors utilized visual word dictionary in combination of Speeded Up Robust Feature SURF descriptor for the feature extraction and multiple classifier were used for the classification of different lesions.…”

Section: Classification Of Microaneurysm and Hemorrhagesmentioning

confidence: 99%

“…Jelinek et al [31] employed SURF feature descriptor and visual word dictionary in feature extraction stage while three classifiers: logical OR fused classifier, Majority Vote fused classifier and Meta SVM were used in the classification stage. The method obtained 0.916 AUC for hard exudates and 0.896 for cotton woolen spot respectively.…”

Section: Classification Of Exudates and Cotton Woolen Spotsmentioning

confidence: 99%

Image processing and classification in diabetic retinopathy: A review

Ahmad

Mansoor

Mumtaz

et al. 2014

2014 5th European Workshop on Visual Information Processing (EUVIP)

View full text Add to dashboard Cite

Diabetic retinopathy is one of the disabling microvascular complications of diabetes mellitus that causes the loss of central vision or in cases complete vision loss if not recognized and cured at the earlier stage. This work reviews the latest techniques in digital image processing and pattern classification employed for the detection of diabetic retinopathy and compares them on the basis of different performance measures like sensitivity, specificity, accuracy and area under the curve in receiver operating characteristic. The classification of diabetic retinopathy follows various steps like pre-processing, feature extraction and classification of microaneurysm, hemorrhages, exudates and cotton woolen spot. In this paper, the reported literature in each domain is analyzed.

show abstract

“…The GMM (Gaussian Mixture Model) classifier is used and the accuracy of 97.3% is achieved. The data fusion method with a meta-SVM classifier for DR detection is implemented in [12]. An automatic method to detect MAs from retinal images using the C4.5 algorithm is implemented in [13].…”

Section: Introductionmentioning

confidence: 99%

Automated Binary and Multiclass Classification of Diabetic Retinopathy Using Haralick and Multiresolution Features

et al. 2020

View full text Add to dashboard Cite

Diabetic Retinopathy (DR) is considered as the complication of Diabetes Mellitus that damages the blood vessels in the retina. This is characterized as a serious vision-threatening problem in most of the diabetic subjects. Effective automatic classification of diabetic retinopathy is a challenging task in the medical field. The feature extraction plays an eminent role in the effective classification of disease. The proposed work focuses on the extraction of Haralick and Anisotropic Dual-Tree Complex Wavelet Transform (ADTCWT) features that can perform reliable DR classification from retinal fundus images. The Haralick features are based on second-order statistics and ADTCWT reliably extracts the directional features in images. The proposed work concentrates on both binary classification as well as multiclass classification of DR. The system is evaluated across various classifiers such as Support Vector Machine (SVM), Random Forest, Random Tree, J48 classifiers by giving input image features extracted from the MESSIDOR, KAGGLE and DIARETDB0 databases. The performances of the classifiers are analyzed by comparing specificity, precision, recall, False Positive Rate (FPR) and accuracy values for each classifier. The evaluation results show that by applying the proposed feature extraction method, Random Forest outperforms all the other classifiers with an average accuracy of 99.7% and 99.82% for binary and multiclass classification respectively. INDEX TERMS DR binary classification, DR multiclass classification, retinal fundus images, HARALICK, ADTCWT, 10-fold cross validation. S. GAYATHRI received the B.Tech. degree in electronics and communication engineering from the Vivekandha College of Engineering, under Anna university, Chennai, in 2012, and the M.Tech. degree in advanced communication and information systems from the Rajiv Gandhi Institute of Technology, under Mahatma Gandhi University, Kottayam, in 2014. She is currently pursuing the Ph.D. degree in biomedical image processing with the National Institute of Technology at Tiruchirappalli, Tiruchirappalli, India. Her research interests include bio medical signal and image processing, machine learning, and deep learning.

show abstract

Data fusion for multi-lesion Diabetic Retinopathy detection

Abstract: Screening of Diabetic Retinopathy (DR)

Cited by 18 publications

References 15 publications

A Weakly-Supervised Framework for Interpretable Diabetic Retinopathy Detection on Retinal Images

A Weakly-Supervised Framework for Interpretable Diabetic Retinopathy Detection on Retinal Images

Image processing and classification in diabetic retinopathy: A review

Automated Binary and Multiclass Classification of Diabetic Retinopathy Using Haralick and Multiresolution Features

Contact Info

Product

Resources

About