Objectron
Object-centric videos with 3D bounding boxes for understanding everyday objects in real scenes.
About
Objectron focuses on common objects observed from moving cameras. Its object-centric sequences and 3D annotations support experiments that connect image evidence to spatial pose and extent.
A synchronized 2D and 3D view makes annotation quality, pose ambiguity, and difficult viewpoints much easier to assess than a frame-by-frame image browser alone.
Suggested agent workflows
Copy a prompt into your agent of choice. Each workflow starts small, respects the source terms, and makes the result inspectable in Rerun.
Audit a training slice
Fetch Objectron from its official source at https://github.com/google-research-datasets/Objectron. Read and preserve its source terms and citation requirements. Build a reproducible manifest for a representative slice, then load it into Rerun. Inspect the RGB video, 3D boxes, Poses streams for timestamp gaps, missing samples, calibration problems, and annotation inconsistencies. Summarize the findings and save both the validation script and a Rerun recording that makes every issue easy to review.
Build a baseline pipeline
Use Objectron to scaffold a baseline pipeline for 3D detection, Object pose, Tracking. Start with a small deterministic cohort and document the selection criteria. Create the dataset adapter, preprocessing steps, train and validation splits, and one lightweight baseline. Log inputs, labels, predictions, losses, and representative failures to Rerun so I can inspect the full experiment timeline. Include exact setup and reproduction commands.
Find useful failure cohorts
Analyze Objectron and propose a practical cohort taxonomy for 3D detection, Object pose, Tracking. Use available metadata and the RGB video, 3D boxes, Poses signals to identify hard cases, rare conditions, coverage gaps, and possible leakage between splits. Materialize the cohorts as queryable metadata, log examples from each cohort to Rerun, and produce a short report explaining which cohorts should become regression tests. Do not infer labels that are not present in the source data.