AI Paper Atlas

Research map / Point clouds / Segmentation and detection

Image-to-Point Cloud Feature Back-Projection for Multimodal Training of 3D Semantic Segmentation

Jiawei Han, Matteo Poggi, Li Huan, Changshuo Wang, Kaiqi Liu, Wei Li

CVPR 2026 · Point clouds › Segmentation and detection · approach: Point cloud semantic segmentation

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train with IPFP back-projected image features (LiDAR-only inference)PTv3 · SemanticKITTI valmIoU: 68.6 -> 71.1Improved
train with IPFP back-projected image featuresPTv3 · nuScenes valmIoU: 80.3 -> 81.4Improved
train with IPFP back-projected image featuresPTv3 · Waymo valmIoU: 71.2 -> 72.4Improved
train with IPFP back-projected image featuresSPVCNN · SemanticKITTI valmIoU: 63.8 -> 65.1Improved
single-branch IPFP instead of dual-branch fusion (PointPainting, 2DPASS)PTv3 · SemanticKITTI valtraining time (s/it): 0.397 -> 0.204Improved
swap depth source (IP-BASIC, Depth Anything V2, Unidepth, UnidepthV2) (ablation)IPFP(PTv3) · SemanticKITTI valmIoU: 70.87 to 71.13No clear effect
increase number of back-projected images per scene (ablation)IPFP(PTv3) · nuScenes valmIoU: 81.0 -> 81.4 (1 -> 3 images)Improved
stochastically discard back-projected image features during training (ablation)IPFP(PTv3) · SemanticKITTI valmIoU: 70.6 -> 71.1Improved
set discard probability too high (gamma 0.7) (ablation)IPFP(PTv3) · SemanticKITTI valmIoU: 71.1 -> 70.5Hurt
too few or too many back-projected cluster centers (2% or 7% vs 5%) (ablation)IPFP(PTv3) · SemanticKITTI valmIoU: 71.1 -> 70.4 (2%), 70.8 (7%)Hurt
use pretrained lightweight CNNs instead of learnable conv+linear for cluster module (ablation)IPFP(PTv3) · nuScenes valmIoU: 81.4 -> 80.7 / 80.6Hurt
raise lower depth percentile (alpha_l) to 50 to drop near-field samples (ablation)IPFP(PTv3) · SemanticKITTI valmIoU: 70.7 (20) -> 71.1 (50) -> 70.9 (60)Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy