Use case: dynamic pipeline execution

[denisri]

Author: Yann C. and Nicolas S.

Summary: Run parallel processes whose number is not known in advance but depends on a pipeline input

Description:

A pipeline will cut an image in fixed size sub-images then launch a process for each sub-image. The result of the sub-image processes will be integrated to generate a result. It could be a very simple processing that would compute the mean of each sub-image to compute a lower resolution image.

Participants: NA => single user

Datasets: A few images with different resolutions. It could be interesting to build a test case to assess the behavior of the platform with very big resolution images such as the ones handled in MIRCen.

Data flow: local image -> platform then platform -> local result

Security: NA

Workgroup(s): pipeline, metadata