Evaluating Image Segmentation Models
A look at evaluation techniques for semantic and instance segmentation.
semantic-composition instance-segmentation computer-vision evaluation natural-language-processing segmentation article tutorial

When evaluating a standard machine learning model, we usually classify our predictions into four categories: true positives, false positives, true negatives, and false negatives. However, for the dense prediction task of image segmentation, it's not immediately clear what counts as a "true positive" and, more generally, how we can evaluate our predictions. In this post, I'll discuss common methods for evaluating both semantic and instance segmentation techniques.

Don't forget to tag @jeremyjordan in your comment, otherwise they may not be notified.

Authors
Machine learning engineer. Broadly curious. Twitter: @jeremyjordan
Share this project
Similar projects
Distribution Based Compositionality Assessment (DBCA)
A method of systematically generating datasets with train and test splits diverging in a controllable and measurable way.
Top collections