MARSAD / CONSOLIDATED CATALOG / #012

Two scooters and a backward pickup

Awaiting your 3D scene approval

Two children cross on kick scooters. One continues while the other stops, looks back and retrieves a dropped toy box.

Original description: 前方两名儿童滑着滑板车横穿马路，其中一人停下回头捡物，另一人继续前行

47 sFull continuous rehearsal
3Critical actors and props
941Exported time steps checked
0New Image 2.5 outputs

Loading the full animated scene…

Critical moments

What to verify

  • Verify exactly two children, two kick scooters and one toy box: the children are included in the two scooter actors, never duplicated as separate actors.
  • Evaluate 941 samples at 20 Hz over -7..40 s; maintain the 2.9 m lateral separation of scooter roots and check articulated scooter, child and hand envelopes during both manual pivots.
  • Show Child A braking, looking back, returning south, crouching and lifting the object while Child B continues north.
  • Verify ground contact through curb transitions without inventing a ramp; the child may articulate and hand-lift the lightweight scooter while the composite support root remains grounded.
  • Verify continuous pickup at 5.3..6 s and attachment without a position jump; retain the same toy ID in the hand through the final frame.
  • Keep ego stopped at x=12 until at least 16 s, after both complete footprints and the carried box are on the north sidewalk.
  • Require full-timeline human review before Image 2.5 edits original Image 1; unresolved source-vehicle or curb conflicts block image generation.

Review notes

Reviewed both child scooter riders, the stop and turn, toy pickup and departure. Riders stay attached to their scooters and the toy transfers continuously to the hand. Close-up contact is approximate but the retrieval sequence is readable.

Check action, relative position, facing, posture and visibility. Approximate offsets are acceptable. Sampled visual checks do not prove every contact or continuous collision clearance.

Approval stops here. Please identify this scene as #012 when approving it or requesting revisions. Image 2.5 generation and training remain paused for this new catalog phase.

Rendered inspection views

Event, approach and final views rendered from the exact committed Higgsfield Blender source. These are procedural 3D previews, not generated training photographs.

Two child riders approach the south curb; one small toy box already lies in the road.
-7 s · Two child riders approach the south curb; one small toy box already lies in the road. · REVIEW_WORK_AREA
Both children begin riding north across the roadway.
-3 s · Both children begin riding north across the roadway. · REVIEW_WORK_AREA
The trailing child looks back and brakes while the other continues forward.
0 s · The trailing child looks back and brakes while the other continues forward. · REVIEW_WORK_AREA
The stopped child finishes a manual scooter pivot toward the dropped box.
3 s · The stopped child finishes a manual scooter pivot toward the dropped box. · REVIEW_WORK_AREA
The returning child crouches beside the box within the ego corridor.
4.4 s · The returning child crouches beside the box within the ego corridor. · REVIEW_WORK_AREA
The box is lifted into the child's hand; the other child waits on the north sidewalk.
6 s · The box is lifted into the child's hand; the other child waits on the north sidewalk. · REVIEW_WORK_AREA
The returning child reaches the north sidewalk with the recovered box.
14 s · The returning child reaches the north sidewalk with the recovered box. · REVIEW_WORK_AREA
After both children remain clear, ego may resume subject to the clearance check.
16 s · After both children remain clear, ego may resume subject to the clearance check. · REVIEW_WORK_AREA
Both children and scooters remain on the north sidewalk; the recovered box stays held.
40 s · Both children and scooters remain on the north sidewalk; the recovered box stays held. · REVIEW_WORK_AREA
Original calibrated forward camera at the critical event
4.4 s · Original calibrated forward camera at the critical event · CAM_F0
Contact check: Dropped toy box
4.4 s · Contact check: Dropped toy box · REVIEW_CONTACT
Contact check: Dropped toy box
6 s · Contact check: Dropped toy box · REVIEW_CONTACT

Original NAVSIM camera context

Original NAVSIM CAM_F0
Original NAVSIM anchor · CAM_F0. This is the untouched camera context, not the synthetic scene.
Original NAVSIM CAM_R0
Original NAVSIM anchor · CAM_R0. This is the untouched camera context, not the synthetic scene.

2 conflicting background vehicles were moved to clear holding positions for this counterfactual rehearsal. Original photos remain unchanged. Before later inpainting, any overlap with original-photo traffic must be resolved. Background positions in this 3D rehearsal are not an exact replay of the photographs.

After scene approval

The working pipeline will edit each original NAVSIM camera photograph with Image 2.5. Its same-time Higgsfield view will provide position and posture, annotated with facing direction and current speed. Temporal image requests will cover the full event with a 1 Hz base plus critical moments.

Assumptions and source provenance

Fresh GPT-6 motion plans use semantic descriptions from the pinned consolidated catalog and NAVSIM environments. No upstream nuPlan vectors, trajectories or images are used. The four source environments are reused; 82 scenario descriptions do not mean 82 independent locations.

Higgsfield revision 1. The portable GLB combines export clips without changing geometry or animation samples.

This review uses Higgsfield revision 1 and the exact exported scene snapshot.

Higgsfield workspace (latest revision) · Download animated GLB · GPT-6 scene plan · Visual review evidence · Motion checks · Approval status