Segmentation

Note. This tutorial's interactive widgets and its citation previews require JavaScript to be enabled in your browser. If you were to ask seven radiologists to outline the same structure on the same medical image, it is almost a certainty that you will get seven different masks. To someone unfamiliar with medical imaging, this might seem like carelessness, but it is not. Boundaries in medical images are genuinely ambiguous, and if you add personal annotation preferences or “style”, expertise levels, and other factors, the disagreement starts to make sense. Across organs and modalities, the volume of a manually contoured structure routinely varies by tens of percent between observers (for example, inter-observer volume differences of 20% and more are common in CT organ delineation), and that variation propagates directly into downstream tasks: radiotherapy decisions, measurements, and the “ground truth” on which segmentation models are trained and evaluated. When these annotations feed a supervised model, the choice of how to collapse them into a single training target is a modeling decision, and this tutorial shows how different methods can make that decision differently. ...

nnU-Net is now considered a standard and state-of-the-art tool for medical image segmentation, but, I think that it is opinionated in the following ways that affect how I may use it: It enforces a specific training and evaluation workflow, including assumptions about data splits. This means using custom train-valid-test splits requires some workarounds. It assumes 3D volumetric data, and while 2D data is supported, it’s the not the primary use case. It assumes grayscale images, and does not support RGB images out of the box. This again means that RGB images require some workarounds. A lot of researchers, however, happily use nnU-Net for 3D volumetric medical images, so these limitations may not be relevant to them. ...

Here are some slides I made to present this MIDL 2021 paper in our reading group:

Just like the last term, I decided to present on a topic spanning 2-3 papers instead of a single paper. This time, I chose to present on how existing works “aggregate” multi-annotator segmentations for medical images. These are the papers that I covered in this presentation: Warfield et al., “Simultaneous Truth and Performance Level Estimation (STAPLE): An Algorithm for the Validation of Image Segmentation”, IEEE Transactions on Medical Imaging, 2004 [URL]. Kats et al., “A Soft STAPLE Algorithm Combined with Anatomical Knowledge”, MICCAI 2019 [URL]. Zhang et al., “Learning to Segment When Experts Disagree”, MICCAI 2020 [URL]. Here are the slides I made to present this topic in our reading group: ...

This term, I really wanted to present this one ICASSP paper I found very interesting, but then I realized that the authors followed up their work with another journal paper (IEEE TIP), so I decided to discuss both. Both of these papers deal with the topic of unsupervised image segmentation: Kanezaki, “Unsupervised Image Segmentation by Backpropagation”, ICASSP 2018 [URL]. Kim et al. “Unsupervised Learning of Image Segmentation Based on Differentiable Feature Clustering”, IEEE Transactions on Image Processing, 2020 [URL]. Here are the slides I made to present this topic in our reading group: ...