Visual explanations for the detection of diabetic retinopathy from retinal fundus images
Boreiko, V.; Ilanchezian, I.; Ayhan, M.; Müller, S.; Koch, L. M.; Faber, H.; Berens, P.; Hein, M.
Show abstract
In medical image classification tasks like the detection of diabetic retinopathy from retinal fundus images, it is highly desirable to get visual explanations for the decisions of black-box deep neural networks (DNNs). However, gradient-based saliency methods often fail to highlight the diseased image regions reliably. On the other hand, adversarially robust models have more interpretable gradients than plain models but suffer typically from a significant drop in accuracy, which is unacceptable for clinical practice. Here, we show that one can get the best of both worlds by ensembling a plain and an adversarially robust model: maintaining high accuracy but having improved visual explanations. Also, our ensemble produces meaningful visual counterfactuals which are complementary to existing saliency-based techniques. Code is available under https://github.com/valentyn1boreiko/Fundus_VCEs.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Clinical Validation of Saliency Maps for Understanding Deep Neural Networks in Ophthalmology 96%
- STAMP: Simultaneous Training and Model Pruning for Low Data Regimes in Medical Image Segmentation 94%
- A Framework for Falsifiable Explanations of Machine Learning Models with an Application in Computational Pathology 93%
Similar papers in this journal
- MultiHeadGAN: A Deep Learning Method for Low Contrast Retinal Pigment Epithelium Cells Segmentation in Fluorescent Flatmount Microscopy Images 96%
- BenchXAI: Comprehensive Benchmarking of Post-hoc Explainable AI Methods on Multi-Modal Biomedical Data 96%
- Deep Multimodal Graph-Based Network for Survival Prediction from Highly Multiplexed Images and Patient Variables 92%
Similar papers in this journal
- Detection and measurement of butterfly eyespot and spot patterns using convolutional neural networks 94%
- Enhancing Breast Ultrasound Segmentation through Fine-tuning and Optimization Techniques: Sharp Attention UNet 93%
- DMENet: Diabetic Macular Edema Diagnosis using Hierarchical Ensemble of CNNs 93%
Similar papers in this journal
Similar papers in this journal
- An Inherently Interpretable AI model improves Screening Speed and Accuracy for Early Diabetic Retinopathy 96%
- Self-supervised contrastive learning improves machine learning discrimination of full thickness macular holes from epiretinal membranes in retinal OCT scans 95%
- Uncovering the effects of model initialization on deep model generalization: A study with adult and pediatric chest X-ray images 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.