AI Paper Atlas

Research map / Point clouds / Point cloud classification

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

Dingkang Liang, Tianrui Feng, Xin Zhou, Yumeng Zhang, Zhikang Zou, Xiang Bai

TPAMI 2025 · Point clouds › Point cloud classification · approach: Self-supervised point cloud pretraining · 48 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
replace full fine-tuning with PointGST spectral tuningPointGPT-L · ScanObjectNN PB_T50_RSoverall accuracy: 93.4 -> 94.83Improved
replace full fine-tuning with PointGSTPoint-BERT · ScanObjectNN OBJ_BGoverall accuracy: 87.43 -> 91.39Improved
replace full fine-tuning with PointGSTPoint-MAE · ScanObjectNN PB_T50_RSoverall accuracy: 85.18 -> 85.29No clear effect
replace full fine-tuning with PointGSTReCon · ScanObjectNN PB_T50_RSoverall accuracy: 90.01 -> 89.49Hurt
use DAPT instead of full fine-tuningACT · ScanObjectNN PB_T50_RSoverall accuracy: 88.21 -> 87.54Hurt
use Point-PEFT instead of full fine-tuningACT · ScanObjectNN OBJ_BGoverall accuracy: 93.29 -> 90.36Hurt
add voting at test timePointGST on PointGPT-L · ScanObjectNN OBJ_BGoverall accuracy: 98.97 -> 99.48Improved
PointGST with strong rotation augmentation vs Point-PEFTPoint-MAE · ScanObjectNN PB_T50_RSoverall accuracy: 89.1 -> 89.3Improved
train on only 2% of ScanObjectNN PB_T50_RS data with PointGST vs other PEFTPointGPT-L · ScanObjectNN PB_T50_RS few-shotoverall accuracy: 81.96% outperforming IDPT, Point-PEFT, and DAPT by 1.74%, 4.34%, and 4.51%Improved
use PointGST instead of DAPTACT · S3DIS Area 5mIoU: 54.5 -> 57.4Improved
use PointGST instead of linear probingPoint-BERT · S3DIS Area 5mIoU: 53.9 -> 58.4Improved
use PointGST instead of full fine-tuningPointGPT-L · ShapeNetPartclass mIoU: 84.80 -> 83.91Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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