Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Per-class saliency maps for M-RCNN #2227

Merged

Conversation

negvet
Copy link
Contributor

@negvet negvet commented Jun 7, 2023

Summary

Generate per-class saliency maps for maskrcnn model, by aggregating raw segmentation masks.

For OV IR model maps are getting generated from filtered raw model predictions of mask head (implemented as postprocessing in OTXMaskRCNNModel wrapper).
For Torch model maps are getting from unfiltered raw model predictions of mask head (implemented as RecordingForwardHook).

Resolves CVS-111726.

To be done in a separate PR: add e2e tests for maskrcnn

Some results (OV iR model, coco dataset):
image

How to test

Checklist

  • I have added unit tests to cover my changes.​
  • I have added integration tests to cover my changes.​
  • I have added e2e tests for validation.
  • I have added the description of my changes into CHANGELOG in my target branch (e.g., CHANGELOG in develop).​
  • I have updated the documentation in my target branch accordingly (e.g., documentation in develop).
  • I have linked related issues.

License

  • I submit my code changes under the same Apache License that covers the project.
    Feel free to contact the maintainers if that's a concern.
  • I have updated the license header for each file (see an example below).
# Copyright (C) 2023 Intel Corporation
# SPDX-License-Identifier: Apache-2.0

@negvet negvet requested a review from a team as a code owner June 7, 2023 15:56
@github-actions github-actions bot added ALGO Any changes in OTX Algo Tasks implementation API Any changes in OTX API labels Jun 7, 2023
@negvet negvet force-pushed the et/per_class_sal_maps_maskrcnn branch from 6133b4c to 4cc1158 Compare June 7, 2023 16:13
@github-actions github-actions bot added the TEST Any changes in tests label Jun 12, 2023
@negvet negvet requested review from eunwoosh and sungmanc June 13, 2023 13:32
Copy link
Contributor

@sovrasov sovrasov left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Overall LGTM, though a bit hacky

@sovrasov sovrasov merged commit cb13b0c into openvinotoolkit:develop Jun 16, 2023
12 checks passed
@negvet negvet mentioned this pull request Jun 16, 2023
8 tasks
sungmanc pushed a commit to sungmanc/training_extensions that referenced this pull request Jun 29, 2023
* per class sal maps for maskrcnn

* tiling support + test enablement/fix

* resolve comments

* enable xai detection e2e tests

* comments + mrcnn xai api test

* nested lists

* linters, sorry for spamming :)
sungmanc added a commit that referenced this pull request Jun 29, 2023
* Per-class saliency maps for M-RCNN (#2227)

* per class sal maps for maskrcnn

* tiling support + test enablement/fix

* resolve comments

* enable xai detection e2e tests

* comments + mrcnn xai api test

* nested lists

* linters, sorry for spamming :)

* Bump-up release version 1.3.1 --> 1.3.2

* Update Changelog

---------

Co-authored-by: Evgeny Tsykunov <evgeny.tsykunov@intel.com>
sungmanc pushed a commit to sungmanc/training_extensions that referenced this pull request Jul 5, 2023
* per class sal maps for maskrcnn

* tiling support + test enablement/fix

* resolve comments

* enable xai detection e2e tests

* comments + mrcnn xai api test

* nested lists

* linters, sorry for spamming :)
@negvet negvet mentioned this pull request Jul 5, 2023
8 tasks
@negvet negvet deleted the et/per_class_sal_maps_maskrcnn branch July 18, 2023 14:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
ALGO Any changes in OTX Algo Tasks implementation API Any changes in OTX API TEST Any changes in tests
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

6 participants