Computer Vision News - November 2019

3 Summary Christian Baumgartner 21 and widely discussed topic in the community with many appl ications. I f our vision as a community is to create ful ly automated systems, it ’s very important that we have quanti f ication of uncertainty at di fferent steps so that error propagation can be control led. On the one hand it has appl ications in automation of steps, but it also has appl ications in directly informing cl inical practitioners about the certainty or uncertainty of an output from a machine learning algorithm.” In terms of next steps, Christian says they are currently working on a journal extension. There are some intriguing properties of the method that haven’t been discussed in the MICCAI paper and they want to analyse further. One of them is that it ’s extremely robust to noise in the images. They bel ieve this is because it has the properties of a shape model as wel l , which they would l ike to investigate further. Best of MICCAI