Looking into Concept Explanation Methods for Diabetic Retinopathy Classification

Andrea M. Storås1,2Orcid, Josefine V. Sundgaard2,3Orcid
1: Department of Holistic Systems, Simula Metropolitan Center for Digital Engineering, Oslo, Norway, 2: Novo Nordisk A/S, Søborg, Denmark, 3: Department of Applied Mathematics and Computer Science, Technical University of Denmark, Kongens Lyngby, Denmark
Publication date: 2024/10/01
https://doi.org/10.59275/j.melba.2024-e7fd
PDF · Code

Abstract

Diabetic retinopathy is a common complication of diabetes, and monitoring the progression of retinal abnormalities using fundus imaging is crucial. Because the images must be interpreted by a medical expert, it is infeasible to screen all individuals with diabetes for diabetic retinopathy. Deep learning has shown impressive results for automatic analysis and grading of fundus images. One drawback is, however, the lack of interpretability, which hampers the implementation of such systems in the clinic. Explainable artificial intelligence methods can be applied to explain the deep neural networks. Explanations based on concepts have shown to be intuitive for humans to understand, but have not yet been explored in detail for diabetic retinopathy grading. This work investigates and compares two concept-based explanation techniques for explaining deep neural networks developed for automatic diagnosis of diabetic retinopathy: Quantitative Testing with Concept Activation Vectors and Concept Bottleneck Models. We found that both methods have strengths and weaknesses, and choice of method should take the available data and the end user’s preferences into account. Our code is available at https://github.com/AndreaStoraas/ConceptExplanations_DR_grading

Keywords

Explainable Artificial Intelligence · Concept-Based Explanations · Diabetic Retinopathy · Fundus Images

Bibtex @article{melba:2024:021:storås, title = "Looking into Concept Explanation Methods for Diabetic Retinopathy Classification", author = "Storås, Andrea M. and Sundgaard, Josefine V.", journal = "Machine Learning for Biomedical Imaging", volume = "2", issue = "iMIMIC 2023 special issue", year = "2024", pages = "2053--2066", issn = "2766-905X", doi = "https://doi.org/10.59275/j.melba.2024-e7fd", url = "https://melba-journal.org/2024:021" }
RISTY - JOUR AU - Storås, Andrea M. AU - Sundgaard, Josefine V. PY - 2024 TI - Looking into Concept Explanation Methods for Diabetic Retinopathy Classification T2 - Machine Learning for Biomedical Imaging VL - 2 IS - iMIMIC 2023 special issue SP - 2053 EP - 2066 SN - 2766-905X DO - https://doi.org/10.59275/j.melba.2024-e7fd UR - https://melba-journal.org/2024:021 ER -

2024:021 cover