AI Paper Atlas

Research map / Point clouds / Segmentation and detection

2D-3D Interlaced Transformer for Point Cloud Segmentation with Scene-Level Supervision

Cheng-Kun Yang, Min-Hung Chen, Yung-Yu Chuang, Yen-Yu Lin

ICCV 2023 · Point clouds › Segmentation and detection · approach: Point cloud semantic segmentation · 23 citations (OpenAlex)

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
add 2D multi-view images through interlaced 2D-3D decoder (ablation)MIT · ScanNet valmIoU: 31.6 -> 35.8Improved
add multi-class token transformer encoder (ablation)3D CAM baseline · ScanNet valmIoU: 26.1 -> 31.6Improved
use interlaced decoder instead of standard cross-attention decoder (ablation)MIT · ScanNet valmIoU: 33.7 -> 35.4Improved
add contrastive N-pair loss on class tokens (ablation)MIT · ScanNet valmIoU: 35.4 -> 35.8Improved
swap query and key-value roles (2D query first) (ablation)MIT interlaced decoder · ScanNet valmIoU: 35.4 -> 35.2No clear effect
use interlaced decoder rather than MLP or BPM fusion (ablation)MIT · ScanNet valmIoU: 32.6 (MLP) / 32.4 (BPM) -> 35.8Improved
add camera poses and depth via 3D world coordinate embeddingMIT · ScanNet valmIoU: 35.8 -> 37.1Improved
use BPM fusion with pose and depth instead of MLP fusion (ablation)MIT · ScanNet valmIoU: 32.6 -> 32.4No clear effect
increase number of views beyond 16 (ablation)MIT · ScanNet pseudo labelsmIoU: 32.7 (16 views) -> 30.9 (32), 31.2 (64)No clear effect
stack more than two interlaced blocks (ablation)MIT · ScanNet pseudo labelsmIoU: 32.7 (R=2) -> 32.1 (R=3), 32.4 (R=4)No clear effect
remove ImageNet pre-training for 2D ResNet-50 (ablation)MIT · ScanNet valmIoU: 35.8 -> 34.6Hurt
flatten image features into patch tokens instead of global average pooling (ablation)MIT · ScanNet train (joint 2D-3D)mIoU: 0.275 (3D) and 0.129 (2D)Hurt

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