Steps to reproduce:
- Load an image
- Load the same image again
- Group the two point data
- Compute any derived data (with at least two dimensions) based on the grouped data (e.g. analyze with PCA, t-SNE, HSNE, ...)
- Open one image
- Drag and drop the derived data on the Image Viewer
Expected outcome:
- Like when doing the same with embeddings of not-grouped data, we would like to automatically create a derived image and re-color the original image based in the embedding positions.
Actual outcome:
Possible solution:
- Disallow accepting grouped point data in the image viewer drop widget
Moments before the crash (we just drag&dropped HSNE Embedding on the image viewer:

Other observations:
- The image dimensions cannot be guessed correctly (it is also not clear what those should be)
- This probably causes a mismatch in number of points.
- When manually setting an image size that results in the same number of points as the grouped data, there is no crash, but the recolored image is not very useful
Top: HSNE with manually set double width and unticking linked data
Bottom: PCA with manually set double width

Steps to reproduce:
Expected outcome:
Actual outcome:
Possible solution:
Moments before the crash (we just drag&dropped

HSNE Embeddingon the image viewer:Other observations:
Top: HSNE with manually set double width and unticking linked data

Bottom: PCA with manually set double width