AUTOMOTIVE / ROBOTICS / CONSTRUCTION / SURVEYING
3D cuboid, 9-DoF tracking, semantic per-point, sensor-fusion alignment for LiDAR + camera + radar. EEA-resident, kappa-gated, calibration-verified.
9-DoF cuboid + tracking EEA-resident
PROCUREMENT READINESS
EU AI Act Article 10 enforcement begins 2 August 2026. Project-specific data-governance evidence supports automotive engagements.
Per-delivery artefact pack
Data governance for high-risk AI in vehicles. Annotation provenance, bias-examination notes, ODD coverage documentation.
Per-contributor consent. Processor agreement. Records of processing. 30-day erasure SLA.
ALKS edge-case taxonomy, project-specific evidence, and EEA residency by default.
We map the artefact pack to your ODD scope, modality mix, and deployment environment.
Which one you procure depends on what your perception stack needs. The double-spend failure mode is ordering cuboid when you needed per-point semantic.
7-DoF bounding box (x, y, z, w, h, l, yaw) plus class. Sufficient for most AD object-detection baselines and counting tasks.
7 floats: x,y,z,w,h,l,yaw + class 3D IoU >= 0.6 vehicle, >= 0.5 pedestrian Adds pitch + roll to the 7-DoF box. Captures orientation for rare poses, robotics grasp planning, and motorcycles or off-road cases.
9 floats: x,y,z,w,h,l,yaw,pitch,roll 3D IoU >= 0.6 + orientation err < 5deg Every LiDAR return labelled with a class. Per-point granularity for road surface, vegetation, drivable area, terrain modelling.
uint8 per point, N classes mIoU >= 0.65 over taxonomy Unified per-point class plus per-instance IDs. Single representation for full scene understanding (stuff and things).
class + instance_id per point PQ >= 0.55, RQ >= 0.7 Persistent instance IDs across frames. Enables multi-frame perception, prediction, and planning. Identity-switch rate reported in delivery.
instance_id stable across t ID-switches < 2% / km Different sensors, different ontologies, same delivery contract: EEA-resident annotation, calibration verified, Article 30 records on delivery.
Cuboid + tracking + multi-modal fusion for perception stacks.
9-DoF cuboid + tracking Warehouse SLAM, delivery autonomy, grasp planning.
cuboid + grasp regions BIM-ready 3D, structural surveying, terrain DEM.
semantic per-point + LAS/LAZ Canopy height, biomass, per-tree segmentation, yield.
semantic per-point Rail, pipeline, power line. Change detection, defect localization.
panoptic 3D + change-diff High-density urban + rural maps for HD positioning.
semantic per-point HOW WE LABEL
Calibration verified BEFORE annotation. 3D IoU thresholds documented. Multi-frame consistency checked.
Class taxonomy locked with your team. Calibration data (intrinsic, extrinsic, PTP-sync timestamps) verified before annotation begins.
Deliverable: Versioned schema + calibration report
Class definitions with ODD scope. Occlusion handling, partial-visibility policy, distance-bucket thresholds for ground truth.
Deliverable: Annotation guideline document
Pilot batch on shared subset. 3D IoU between annotators computed. Camera-LiDAR projection check. Disagreement patterns drive refinement.
Deliverable: Calibration 3D-IoU report
Production starts only when calibration clears the task-specific agreement, geometry, distance, and projection thresholds in the acceptance plan.
Deliverable: Gate-pass attestation
Multi-frame interpolation between keyframes. Tracker continuity check. Single-tenant CVAT 3D + proprietary fusion viewer.
Deliverable: Annotated scans + tracking
Per-class 3D IoU. BEV mAP. Tracking continuity. Camera-LiDAR projection consistency. Article 30 records, signed DPA, sub-processor list.
Deliverable: Final delivery + metrics report
Six gates. One trail of evidence. Calibration verified before any annotation.
Six representative driving scenes, each with the annotation class taxonomy we lock with your team before production. Class counts shift per scene type. Production work runs against your raw scans, multi-frame, calibration-verified.
3D IoU at 0.7 is the procurement-grade threshold for autonomous-driving perception. We hold this on the public splits we have published; customer-deployment numbers are reported under NDA.
Most LiDAR public datasets are research-licensed; A2D2 is the notable commercial-clean exception. Production work runs against your own data under your engagement DPA.
| Benchmark | Split | 3D IoU @ 0.7 | Distance buckets | Tracking continuity | License |
|---|---|---|---|---|---|
| KITTI / KITTI-360 KITTI format | Test | in-house: 94% | 0-30, 30-60, 60m+ | < 1.6% switch / km | Research only CC BY-NC-SA |
| nuScenes nuScenes JSON | Val | in-house: 92% | 0-30, 30-60, 60m+ | < 2.1% switch / km | Research only Non-commercial |
| Waymo Open TFRecord, proto | Val | supported | 0-30, 30-50, 50m+ | supported | Research only Waymo dataset |
| A2D2 (Audi) PNG + JSON | Full | in-house: 89% | 0-30, 30-60, 60m+ | < 2.4% switch / km | Commercial OK CC BY-ND 4.0 |
| SemanticKITTI 25 classes/point | Val | mIoU 0.71 | per-point, no buckets | n/a (segmentation) | Research only CC BY-NC-SA 4.0 |
| Argoverse 2 AV2 SDK | Val | supported | 0-30, 30-100m | supported | Mixed terms CC BY-NC-SA 4.0 |
License status reflects publicly stated terms. Verify per engagement before commercial training use. 'in-house' figures are anonymized customer-deployment results under NDA.
WHAT YOU RECEIVE
The records a regulated AD or robotics buyer expects with every 3D engagement. No upgrade tier, no separate request.
Versioned guideline with class definitions, occlusion handling, partial-visibility policy, distance-bucket thresholds. ODD scope documented. Every version preserved for audit.
Intrinsic, extrinsic, and PTP-sync timestamp verification before any annotation begins. Calibration drift checks on long sequences. Camera-LiDAR projection consistency report.
Per-class 3D IoU at multiple thresholds. BEV mAP. Per-distance-bucket accuracy (near, mid, far). Orientation error (yaw, pitch, roll). Center-distance error.
Multi-frame tracking ID continuity. Track switch rate. Per-track length distribution. Identity switches per kilometre or per frame.
Article 30 records of processing, signed Article 28 DPA, lawful-basis documentation, 30-day erasure SLA, and full sub-processor list with Article 28(2) change notifications.
START A PROJECT
Short brief now, deeper scoping in the reply.
Camera intrinsic + extrinsic parameters, LiDAR-camera extrinsics, PTP-synchronized timestamps. Calibration verified before annotation begins; drift checks on long sequences. Calibration verification log shipped with every delivery.
Yes. LiDAR + camera (most common), LiDAR + camera + radar, LiDAR + IMU + GPS, LiDAR + thermal. Single 3D box drawn in LiDAR and projected to all cameras. Time-synced multi-frame annotation.
EEA-resident. Norwegian company, EEA contributor network, EEA infrastructure. 30-day GDPR Article 17 erasure SLA. Norwegian jurisdiction; transfer controls defined per project.
GDPR Article 28 processor obligations, EU AI Act Article 10 data governance, EEA residency by default, DPA-by-default, and project-specific evidence. Records of processing under GDPR Article 30 ship with every delivery.
KITTI format, nuScenes JSON, custom JSON or Protocol Buffers, vendor formats (Ouster, Velodyne, Hesai), PCD, LAS/LAZ for surveying. Customer-owned work product, no reuse rights retained.
One business day reply. NDA on request. DPA included.
Your information is never shared. We respond with the next scoping step.
Add YPAI to your home screen
Tap the Share button, then Add to Home Screen.