CloudAnalyzer Perception Artifact Comparison

This static report compares a geometry-first non-deep baseline and a higher-fidelity deep baseline on the same public RELLIS-3D frame.

Both candidates are deterministic demo artifacts derived from the reference scene for documentation purposes; they are not archived model outputs.

Summary

Referencereference_scene.pcd
Files2
Mean AUC0.8081
Mean Chamfer0.0673
Best AUCcandidates/deep_baseline.pcd (0.9515)
Worst AUCcandidates/nondeep_baseline.pcd (0.6648)
Best Chamfercandidates/deep_baseline.pcd (0.0252)
Worst Chamfercandidates/nondeep_baseline.pcd (0.1094)
Min AUC0.9000
Max Chamfer0.0500
Pass1
Fail1

Results

Path Points Chamfer AUC Best F1 Threshold Status
candidates/deep_baseline.pcd343700.02520.95151.00000.10PASS
candidates/nondeep_baseline.pcd169060.10940.66480.97350.30FAIL

Inspection Commands

Path Inspection
candidates/deep_baseline.pcd
ca web candidates/deep_baseline.pcd reference_scene.pcd --heatmap
ca heatmap3d candidates/deep_baseline.pcd reference_scene.pcd -o deep_baseline_vs_reference_scene_heatmap.png
candidates/nondeep_baseline.pcd
ca web candidates/nondeep_baseline.pcd reference_scene.pcd --heatmap
ca heatmap3d candidates/nondeep_baseline.pcd reference_scene.pcd -o nondeep_baseline_vs_reference_scene_heatmap.png