Skip to content

[codex] Harden TrackNetV3 pixel detection#102

Draft
Dhruva05 wants to merge 2 commits into
mainfrom
objectDetection
Draft

[codex] Harden TrackNetV3 pixel detection#102
Dhruva05 wants to merge 2 commits into
mainfrom
objectDetection

Conversation

@Dhruva05

@Dhruva05 Dhruva05 commented Jul 3, 2026

Copy link
Copy Markdown

Summary

  • Fix TrackNetV3 subtract-mode preprocessing so background differences clip before uint8 conversion instead of wrapping.
  • Add focused TrackNetV3 tests for heatmap decoding, confidence thresholding, subtract preprocessing, launch/config surface, and node pixel/visibility publishing.
  • Document the v1 2D pixel-detection scope, checkpoint mounting, upstream repo boundary, and PyTorch dependency strategy.

Validation

  • PYTHONPATH=autonomy/tracknetv3 /Users/dhruvashekdar/.cache/codex-runtimes/codex-primary-runtime/dependencies/python/bin/python3 -m unittest discover autonomy/tracknetv3/test -v passed: 11 tests.
  • PYTHONPATH=autonomy/tracknetv3 python3 -m unittest discover autonomy/tracknetv3/test -v passed: 9 tests, 2 skipped because the system Python lacks NumPy/Pillow.
  • python3 -m compileall -q autonomy/tracknetv3/tracknetv3 autonomy/tracknetv3/launch passed.
  • git diff --check passed.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant