AI Paper Atlas

Research map / Point clouds / Segmentation and detection

4D-Net for Learned Multi-Modal Alignment

AJ Piergiovanni, Vincent Casser, Michael S. Ryoo, Anelia Angelova

ICCV 2021 · Point clouds › Segmentation and detection · approach: Multimodal and open-vocabulary 3D detection · 68 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 dynamic connections (ablation)4D-Net · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 73.1 -> 73.6Improved
add connectivity search (ablation)4D-Net (no dynamic) · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 72.6 -> 73.1Improved
add projection fusion (ablation)4D-Net (no conn, no dyn) · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 62.5 -> 72.6Improved
add RGB in time (video) instead of single RGB frame (ablation)4D-Net · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 72.6 -> 73.1Improved
add point clouds in time to single PC (ablation)PointPillars · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 55.7 -> 60.5Improved
add a single RGB image with projection and connection learning (ablation)single PC · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 55.7 -> 64.3Improved
increase point clouds from 16 to 32 (ablation)PointPillars PC-only · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 60.5 -> 60.3No clear effect
increase RGB resolution from 224 to 512 (ablation)single PC + RGB · Waymo Open Dataset val (vehicles, 0.7 IoU)AP 50m+: 47.2 -> 52.4Improved
use video input instead of single image at 224x224 (ablation)single PC + RGB · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 64.3 -> 65.4Improved
replace projection with spatial average pooling (ablation)single PC + RGB · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 64.3 -> 58.9Hurt
replace projection with spatial transformer (ablation)single PC + RGB · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 64.3 -> 59.5Hurt
replace Tiny Video Network with AssembleNet (ablation)single PC + RGB · Waymo Open Dataset val (vehicles, 0.7 IoU)AP L1: 65.4 -> 66.8No clear effect

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