Papers with Code · CC BY-SA 4.0

3D Point Cloud Classification

3D Point Cloud Reconstruction › 3D Point Cloud Classification 413 평가 기록 23 datasets

개요는 각 항목의 일부입니다. 위 탭에서 전체 기록을 페이지별로 탐색할 수 있습니다.

벤치마크

Papers with Code 아카이브의 데이터셋별 기록 예시입니다. 평가 조건과 지표 방향이 검증되지 않아 순위를 의미하지 않습니다.

Dataset기록된 모델 예시MetricPaperCode

ModelNet40

PointGST

Overall Accuracy: 95.3

논문 코드

ScanObjectNN

OmniVec2

Overall Accuracy: 97.2

논문 -

ModelNet40 10-way (10-shot)

Point-JEPA

Overall Accuracy: 95.0Standard Deviation: 3.6

논문 코드

ModelNet40 10-way (20-shot)

ReCon++

Overall Accuracy: 96.5Standard Deviation: 3.0

논문 코드

ModelNet40 5-way (10-shot)

PointGPT

Overall Accuracy: 98.0Standard Deviation: 1.9

논문 코드

ModelNet40 5-way (20-shot)

ReCon++

Overall Accuracy: 99.5Standard Deviation: 0.8

논문 코드

ModelNet40

Uni3D

Accuracy (%): 88.2

논문 코드

ModelNet40-C

OmniVec2

Error Rate: 0.142

논문 -

IntrA

3DMedPT

F1 score (5-fold): 0.936

논문 코드

ScanObjectNN

Mamba3D

Overall Accuracy (PB_T50_RS): 92.64GFLOPs: 3.9Number of params (M): 16.9

논문 코드

ScanObjectNN

ReCon++

OBJ_ONLY Accuracy(%): 65.4

논문 코드

ModelNet40

Ours

Classification Accuracy: 93.6

논문 코드

데이터셋

23개

ModelNet40 ScanObjectNN ModelNet40 10-way (10-shot)ModelNet40 10-way (20-shot)ModelNet40 5-way (10-shot)ModelNet40 5-way (20-shot)ModelNet40 ModelNet40-C IntrA ScanObjectNN ScanObjectNN ModelNet40

논문

GPSFormer: A Global Perception and Local Structure Fitting-based Transformer for Point Cloud Understanding

2407.13519 2024-07-18 이 페이지의 평가 2건

Overall Accuracy: 95.4Mean Accuracy: 93.8Number of params: 2.36MFLOPs: 0.7G

Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Model

2404.14966 2024-04-23 이 페이지의 평가 2건

Overall Accuracy: 92.64Number of params: 16.9MFLOPs: 3.9GOBJ-BG (OA): 94.49

Asymmetric Dual Self-Distillation for 3D Self-Supervised Representation Learning

2506.21724 2025-06-26 이 페이지의 평가 1건

Overall Accuracy: 93.72OBJ-BG (OA): 96.73OBJ-ONLY (OA): 94.32

Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining

2302.02318 2023-02-05 이 페이지의 평가 1건

Overall Accuracy: 91.26OBJ-BG (OA): 95.35OBJ-ONLY (OA): 93.80

OmniVec: Learning robust representations with cross modal sharing

2311.05709 2023-11-07 이 페이지의 평가 1건

Overall Accuracy: 96.1

OmniVec2 - A Novel Transformer based Network for Large Scale Multimodal and Multitask Learning

2024-01-01 이 페이지의 평가 1건

Overall Accuracy: 97.2

Parameter-Efficient Fine-Tuning in Spectral Domain for Point Cloud Learning

2410.08114 2024-10-10 이 페이지의 평가 1건

Overall Accuracy: 96.18OBJ-BG (OA): 99.48OBJ-ONLY (OA): 97.76

PointGPT: Auto-regressively Generative Pre-training from Point Clouds

2305.11487 2023-05-19 이 페이지의 평가 1건

Overall Accuracy: 93.4OBJ-BG (OA): 97.2OBJ-ONLY (OA): 96.6

ShapeLLM: Universal 3D Object Understanding for Embodied Interaction

2402.17766 2024-02-27 이 페이지의 평가 1건

Overall Accuracy: 95.25OBJ-BG (OA): 98.80OBJ-ONLY (OA): 97.59

ULIP-2: Towards Scalable Multimodal Pre-training for 3D Understanding

2305.08275 2023-05-14 이 페이지의 평가 1건

Overall Accuracy: 91.5Mean Accuracy: 91.2Number of params: 1.4M

라이브러리

Papers with Code 리더보드 row에 연결된 구현 저장소입니다.

github.com/qizekun/ReCon 2 논문

github.com/runpeidong/act 2 논문

github.com/changshuowang/GPSFormer 1 논문

github.com/xhanxu/Mamba3D 1 논문

github.com/aHapBean/PCP-MAE 1 논문

github.com/asterisci/point-gcc 1 논문

github.com/CGuangyan-BIT/PointGPT 1 논문

github.com/jerryfeng2003/pointgst 1 논문

하위 태스크

F3 Few-Shot 3D Point Cloud Classification 8 데이터셋 · 128 rows ZT Zero-Shot Transfer 3D Point Cloud Classification 3 데이터셋 · 30 rows 3O 3D Object Classification 6 데이터셋 · 27 rows SO Supervised Only 3D Point Cloud Classification 1 데이터셋 · 12 rows