How Reliable Are Out-of-Distribution Generalization Methods for Medical Image Segmentation?

Sanner, Antoine; González, Camila; Mukhopadhyay, Anirban

doi:10.1007/978-3-030-92659-5_39

Cited by 16 publications

(8 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While these heatmaps showed that the authors' models only focused on the part of the input image containing the cancerous cell, the heatmaps were not precise enough to show which parts of the cell were important. Furthermore, DL models are great at learning statistical biases in the data, which may be specific to certain acquisition protocols or hardware 9 . This in turn can cause a decrease in generalizability of the model and decreased performances on external test data, as could also be observed in this study.…”

mentioning

confidence: 54%

Editorial for “Deep Learning With an Attention Mechanism for Differentiating the Origin of Brain Metastasis Using MR Images”

Sanner

Othman

2023

Magnetic Resonance Imaging

Self Cite

View full text Add to dashboard Cite

mentioning

confidence: 54%

Editorial for “Deep Learning With an Attention Mechanism for Differentiating the Origin of Brain Metastasis Using MR Images”

Sanner

Othman

2023

Magnetic Resonance Imaging

Self Cite

View full text Add to dashboard Cite

“…This is denoted as silent failure . Allowing for model adaptation also makes it possible for the model to work in new hospital environments 9 10 16 .…”

Section: Methodsmentioning

confidence: 99%

“…After the radiologist completes the report, the new ground truth annotation can be fed back to the AI model for further improvement. This functionality opens up the possibility of adapting the AI system over time, which prevents the deterioration of model performance as time passes and data distribution changes [9,10]. This loss in performance often goes unnoticed, as deep learning models report high confidence even for low-quality predictions.…”

Section: Integration and Workflow For Lifelong Learningmentioning

confidence: 99%

Closing the loop for AI-ready radiology

Fuchs,

Gonzalez,

Frisch

et al. 2023

Rofo

Self Cite

View full text Add to dashboard Cite

Background In recent years, AI has made significant advancements in medical diagnosis and prognosis. However, the incorporation of AI into clinical practice is still challenging and under-appreciated. We aim to demonstrate a possible vertical integration approach to close the loop for AI-ready radiology. Method This study highlights the importance of two-way communication for AI-assisted radiology. As a key part of the methodology, it demonstrates the integration of AI systems into clinical practice with structured reports and AI visualization, giving more insight into the AI system. By integrating cooperative lifelong learning into the AI system, we ensure the long-term effectiveness of the AI system, while keeping the radiologist in the loop. Results We demonstrate the use of lifelong learning for AI systems by incorporating AI visualization and structured reports. We evaluate Memory Aware-Synapses and Rehearsal approach and find that both approaches work in practice. Furthermore, we see the advantage of lifelong learning algorithms that do not require the storing or maintaining of samples from previous datasets. Conclusion In conclusion, incorporating AI into the clinical routine of radiology requires a two-way communication approach and seamless integration of the AI system, which we achieve with structured reports and visualization of the insight gained by the model. Closing the loop for radiology leads to successful integration, enabling lifelong learning for the AI system, which is crucial for sustainable long-term performance. Key Points:

show abstract

“…We hereby present MiniVess, an expert-annotated dataset of 70 3D 2PFM image volumes of rodent cerebrovasculature. The dataset can be used for training segmentation networks 38,39 , fine-tuning self-supervised pre-trained networks >31,>32,>40 , and as an external validation set for assessing a model’s generalizability 41 . The 3D volumes in this dataset have been curated to only contain clean XYZ imaging in order to ensure correct and consistent annotations, or segmentations, which has been observed to be integral to the evaluation of machine learning models 42 .…”

Section: Background and Summarymentioning

confidence: 99%

MiniVess: A dataset of rodent cerebrovasculature from in vivo multiphoton fluorescence microscopy imaging

Poon

Teikari

Rachmadi

et al. 2022

Preprint

View full text Add to dashboard Cite

We present MiniVess, the first annotated dataset of rodent cerebrovasculature, acquired using two-photon fluorescence microscopy. MiniVess consists of 70 3D image volumes with segmented ground truths. Segmentations were created using traditional image processing operations, a U-Net, and manual proofreading. Code for image preprocessing steps and the U-Net are provided. Supervised machine learning methods have been widely used for automated image processing of biomedical images. While much emphasis has been placed on the development of new network architectures and loss functions, there has been an increased emphasis on the need for publicly available annotated, or segmented, datasets. Annotated datasets are necessary during model training and validation. In particular, datasets that are collected from different labs are necessary to test the generalizability of models. We hope this dataset will be helpful in testing the reliability of machine learning tools for analyzing biomedical images.

show abstract

How Reliable Are Out-of-Distribution Generalization Methods for Medical Image Segmentation?

Cited by 16 publications

References 8 publications

Editorial for “Deep Learning With an Attention Mechanism for Differentiating the Origin of Brain Metastasis Using MR Images”

Editorial for “Deep Learning With an Attention Mechanism for Differentiating the Origin of Brain Metastasis Using MR Images”

Closing the loop for AI-ready radiology

MiniVess: A dataset of rodent cerebrovasculature from in vivo multiphoton fluorescence microscopy imaging

Contact Info

Product

Resources

About