IMViewer

[category:IMViewer]

Browser based user initiated large scale image processing

Viewer interaction

• Browser based user interaction allows following selections using large image viewer
  1. Sub-region at desired resolution (maybe with more general shapes)
  2. Process to be performed on the image and its parameters
  3. The Tag and way to store results
  4. Time of execution

• After user submits the request the request itself is stored in database (so next visitor may see what processing has been configured)

• The results (metadata and the result image-chunks at multiple resolutions
  1. Stored as another document in the MongoDB
  2. Can be viewed along with the base image
  3. Can be shared / reviewed by another user

Server administration

• Maintain a list of possible applications (algorithms) that can be invoked,
  1. Inputs they accept
  2. Error handling and reporting
  3. Estimate of required time and timeout
  4. Resources (RAM, disk space, etc) they will requirA

• Batch process execution
  1. Monitor the progress, see the results
  2. Interrupt the process
     1. Take a look at results
     2. Modify the settings
  3. Resume / discard

• Resource managements
  1. User priviledges
  2. Time scheduling

• Creation of reports

Relation to Midas

Not sure how what is already included in current (or current vision of) Midas, and how existing Midas infrastructure can be leveraged once large image viewer is integrated in Midas.

1. Making Midas a processing powerhouse

Some use cases

1. Crowd sourcing of manual (seed based) segmentation and neuron tracking process (similar to Zooniverse).
2. Manual marking for training data (e.g. cell membrane, mitochondria ..)
3. Remote invocation of anything more involving