AI Paper Atlas

Research map / Point clouds / Point cloud classification

Continuous SO(3) Equivariant Convolution for 3D Point Cloud Analysis

Jaein Kim, HEE BIN YOO, Dong-Sig Han, Yeon-Ji Song, Byoung-Tak Zhang

ECCV 2024 · Point clouds › Point cloud classification · approach: Local feature aggregation networks · 1 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
use continuous SO(3) equivariant filters instead of discrete group convolution when test rotations differ from trainingCSEConv vs E2PN · ModelNet40, train I / test SO(3)accuracy (%): 44.47 -> 83.75Improved
apply SO(3) augmentation to non-equivariant model training instead of nonePCT · ModelNet40, test SO(3)accuracy (%): 15.84 -> 84.20Improved
add SO(3) augmentation during training for aligned test setnon-equivariant PCT · ModelNet40, test Iaccuracy (%): 91.25 -> 86.39Hurt
use continuous-group equivariant filters instead of discretized groupequivariance to random rotations · ModelNet40, no train augmentationaccuracy (%): 44.47 -> 83.75Improved
double the number of KNN neighbors in the integration (ablation)CSEConv · equivariance error on 512 pointsequivariance errorImproved
implement filter as small weight-sharing coordinate network instead of tensor grid parametersCSEConv · ModelNet40 training speedbatches per second: 8.64 -> 27.46Improved
use CSEConv instead of E2PN on real-world scans with aligned train and testScanObjectNN OBJ_ONLY · ScanObjectNN, train I / test Iaccuracy (%): 82.70 -> 72.29Hurt
use CSEConv instead of Frame Averaging, SE3-T and VN on ScanObjectNNcontinuous group equivariant methods · ScanObjectNN, train I / test SO(3)accuracy (%): 71.94 -> 72.29 (vs FA)Improved

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