Traffic Video Vehicle Annotation

I have a library of traffic-camera footage that needs detailed vehicle annotation. Your task is to go through each clip, mark every vehicle frame-by-frame, and track its position and movement across the sequence. No pedestrian or traffic-sign data are required on this project—only vehicles.

The raw material arrives as 1080p MP4s. You may work in CVAT, Labelbox, V7, or another professional tool as long as the final export is delivered in a COCO-style JSON (or compatible CSV) with unique track IDs and time-stamped coordinates. Please keep the original video resolution and frame rate intact so I can align the labels with my downstream computer-vision pipeline.

Deliverables
• Annotated data file per video (JSON or CSV)
• The corresponding project files from your chosen tool
• A brief log noting any frames that were skipped or visually ambiguous

Acceptance criteria
• 100 % of vehicles in every frame are boxed and consistently tracked
• IoU ≥ 0.8 on random spot checks
• File names and time stamps match the originals exactly

This is a fully remote assignment. Let me know your preferred tool and approximate throughput (minutes of video you can label per day) so we can plan the hand-off schedule.