Use volume series as input for heterogeneous refinement or 3D classification


Using “intermediates” mode of 3D-VA display, the resulting volumes are provided as a volume series, as is also the case for 3D-classification.

I would like to be able to use these as input for 3D-classification (beta), in order to test whether the distribution follows the assumptions inherent in 3D-VA, when separated by reference based classification.

I can do this by downloading the volume series, splitting it up and reimporting, but it is rather laborious. It would therefore be useful if volume series could be provided as input to any job that already accepts multiple volume inputs.

Obviously the job would then need to unpack the series and split it up on the backend, but I think this will be increasingly useful in cases where we are performing classifications with large numbers of classes, rather than having to provide them as individual volumes.



Thanks @olibclarke. Discussed internally and duly noted for a future release!