We’re currently working on a new capability for Zegami which adds a new dimension to image analysis, and we’re really excited to be able to share a preview of it.
Zegami is a great way to visualise large sets of images. But what if your images come in pairs, or trios? Or entire stacks? Many users have data where each row references not one but many images. There might be multiple imaging angles per patient, multiple slices for each three-dimensional scan, or perhaps different imaging spectra for a fixed subject.
For these reasons and more, we are adding the ability to include multiple image sources within a collection. Up until now, creating a collection has involved selecting the ‘image’ column from the dataset (which we usually detect automatically). Soon it will be possible to include any number of columns in your data which reference image filenames, and switch between these in the viewer.
This new capability is still in progress, but we have been able to make a preview available in a demonstration collection.
This collection examines chest X-rays, including some of COVID-19 patients. Included in the data are the outputs for each image of an AI model trained to identify COVID-19.
We have previously demonstrated such a collection, but this time we include a secondary image source that shows an explainability map highlighting which regions had the most influence on the verdict of the model. The sources can be switched using the ‘Image Source’ control on the top bar.
Please use the above link to try this feature out for yourselves. We’ll make an announcement when it is possible for customers to create collections with this capability. In the meantime, we’d love to hear from you if you think this might be useful, or if you have any questions or suggestions, so please get in touch!