Skip to content

Add How to Tell Objects Apart with ReID notebook - #450

Open
AlexBodner wants to merge 15 commits into
roboflow:mainfrom
AlexBodner:feat/how-to-tell-objects-apart-with-reid
Open

Add How to Tell Objects Apart with ReID notebook#450
AlexBodner wants to merge 15 commits into
roboflow:mainfrom
AlexBodner:feat/how-to-tell-objects-apart-with-reid

Conversation

@AlexBodner

@AlexBodner AlexBodner commented Jul 30, 2026

Copy link
Copy Markdown
Contributor

Summary

  • Adds a Colab-ready tutorial for the reid package: load a pretrained encoder, reproduce Market-1501 gallery metrics with ReIDEvaluator, and visualize query-to-gallery ranking on SportsMOT soccer crops
  • Uses sports visuals for the ranking demo; Market-1501 is the metrics dataset
  • Registers the notebook in automation/notebooks-table-data.csv and regenerates the README table

PR checklist

  • Release / publish reid docs on reid.roboflow.com (metrics guide at /latest/learn/metrics/) so the notebook doc links resolve
  • After docs are live, spot-check and update notebook URLs if the final docs paths differ from /latest/...
  • Note: SportsMOT crop generation uses a local generate_mot_patches helper because PyPI reid does not ship reid.data.mot yet. After that lands in a published reid release, switch the notebook to from reid.data.mot import generate_mot_patches and drop the inline helper

Test plan

  • Open notebook in Colab / local Jupyter and run all
  • README models table links resolve once merged

AlexBodner and others added 2 commits July 30, 2026 13:08
Copy the reid package eval_reid tutorial into Roboflow notebooks with banner, relative paths, and setup aligned for a first-release showcase.

Co-authored-by: Cursor <cursoragent@cursor.com>
Showcase the reid package with Market/MSMT metrics and SportsMOT ranking visuals, and register it in the notebooks table.

Co-authored-by: Cursor <cursoragent@cursor.com>
@review-notebook-app

Copy link
Copy Markdown

Check out this pull request on  ReviewNB

See visual diffs & provide feedback on Jupyter Notebooks.


Powered by ReviewNB

AlexBodner and others added 13 commits July 30, 2026 13:35
Use trackers.roboflow.com and reid.roboflow.com instead of GitHub blob paths where docs exist.

Co-authored-by: Cursor <cursoragent@cursor.com>
PyPI reid does not ship MOT patch generation yet, so use a local helper for the SportsMOT ranking section.

Co-authored-by: Cursor <cursoragent@cursor.com>
Prefer a local official extract when present; otherwise download the community mirror used by the original eval notebook.

Co-authored-by: Cursor <cursoragent@cursor.com>
One metrics dataset is enough for the tutorial; SportsMOT remains the ranking visualization.

Co-authored-by: Cursor <cursoragent@cursor.com>
Sample random crop pairs and classify same vs different player with an illustrative threshold so the notebook shows ReID beyond gallery ranking.

Co-authored-by: Cursor <cursoragent@cursor.com>
Require a minimum frame gap and prefer larger gaps so pairs are not near-duplicate consecutive frames.

Co-authored-by: Cursor <cursoragent@cursor.com>
Match the step-by-step subsection style used in the other ReID tutorial so each code cell has short context before it.

Co-authored-by: Cursor <cursoragent@cursor.com>
Use the same Setup/Results/closing pattern, shorter imperative section prose, and GitHub package links.

Co-authored-by: Cursor <cursoragent@cursor.com>
0.25 is the BoT-SORT / trackers default appearance threshold; keep paper references for the published MOT tables only.

Co-authored-by: Cursor <cursoragent@cursor.com>
Point the closing link at PR roboflow#449 instead of a main-branch Colab URL that is not published yet.

Co-authored-by: Cursor <cursoragent@cursor.com>
That tutorial belongs on feat/how-to-add-reid-to-trackers; keep this PR focused on the tell-objects notebook.

Co-authored-by: Cursor <cursoragent@cursor.com>
Thin identity crops and filter gallery neighbors by a minimum frame gap so query-to-gallery demos are not dominated by adjacent frames.

Co-authored-by: Cursor <cursoragent@cursor.com>

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Pull request overview

Adds a new Colab-ready tutorial notebook demonstrating object re-identification (ReID) with the reid package, and wires it into the repo’s autogenerated notebook index/table.

Changes:

  • Adds a new tutorial notebook: “How to Tell Objects Apart with ReID”.
  • Registers the notebook in automation/notebooks-table-data.csv so it’s included in the autogenerated README table.
  • Regenerates the README notebooks table to include the new entry and updated count.

Reviewed changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated 1 comment.

File Description
README.md Updates the autogenerated notebooks table (count + new ReID entry).
automation/notebooks-table-data.csv Adds the new notebook metadata row used by the table generator.
notebooks/how-to-tell-objects-apart-with-reid.ipynb New tutorial notebook covering encoder loading, Market-1501 evaluation, and SportsMOT ranking visualization.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment on lines 336 to 352
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.

2 participants