User Story #9948 (new)
Opened 11 years ago
Last modified 11 years ago
Analysis: Feature extraction and storage
Reported by: | spli | Owned by: | spli |
---|---|---|---|
Priority: | critical | Milestone: | Unscheduled |
Component: | General | Keywords: | analysis |
Cc: | analysis@… | Story Points: | n.a. |
Sprint: | n.a. | Importance: | n.a. |
Total Remaining Time: | n.a. | Estimated Remaining Time: | n.a. |
Description (last modified by spli)
This part of the analysis API will cover running and monitoring analysis jobs.
This may include optional methods to access images, datasets, and feature storage so that it's easier for people with little knowledge of the OMERO APIs to integrate their algorithms.
This is particularly the case for feature storage related to machine learning, where it's desirable to have everyone using a common pattern, so as to facilitate interoperability between image feature algorithms and learning/visualisation algorithms.
For an idea of the potential scale that any Alternative Storage will have to cope with (in terms of volume and read/write thoughput) consider full slide pathology scans:
- Scanned slides can have well over 1,000,000 tiles.
- Each tile could have 3,000 features or more.
- Features may not be a plain vector of doubles, each one could be named and typed.
- There could be 1000s of these slides
See also #3519 for a previous requirement of distributed analysis.