AI Paper Atlas

Research map / Detection and segmentation / 3D detection and driving

Finding Your (3D) Center: 3D Object Detection Using a Learned Loss

David Griffiths, Jan Boehm, Tobias Ritschel

ECCV 2020 · Detection and segmentation › 3D detection and driving · approach: Unified LiDAR 3D detection · 10 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
train the scene network with the learned loss on unlabeled scenes instead of sliding the loss networkOurs · S3D chair, 100% labelsmAP@.25: .598 -> .803Improved
train with the learned loss instead of supervised Chamfer lossOurs · S3D chair, 5% labelsmAP@.25: .352 -> .642Improved
train with the learned loss instead of supervised Chamfer lossOurs · S3D chair, 100% labelsmAP@.25: .756 -> .803Improved
train with the learned loss instead of sliding window with very few labelsOurs · S3D chair, 1% labelsmAP@.25: .473 -> .366Hurt
increase label ratio from 1% to 100%Ours · S3D chairmAP@.25: .366 -> .803Improved
increase label ratio from 1% to 100%Ours · S3D chairChamfer error: .554 -> .178Improved
add unlabeled ScanNet scenes to scene network training (loss net from S3D)Ours · ScanNet test, 5% S3D labelsmAP@.25: .138 -> .599Improved
increase number of unlabeled ScanNet training scenesOurs · ScanNetmAP@.25: .079 -> .599Improved
increase loss network labels from 1% to 100%Loss network · S3D chairaccuracy: .853 -> .960Improved
use learned loss instead of supervised or sliding on other classesOurs · S3D bookcase, 5% labelsmAP@.25: .370 -> .640Improved
reduce labels from 100% to 5% on ScanNet V2Ours · ScanNet V2mAP@.25 (%): 50.2 -> 48.3No clear effect
add overlap avoidance (winner-takes-all collision penalty) during scene network training (ablation)Ours · scene network trainingdetectionImproved

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