AI Paper Atlas

Research map / Point clouds / Point cloud classification

Instance-aware Dynamic Prompt Tuning for Pre-trained Point Cloud Models

Yaohua Zha, Jinpeng Wang, Tao Dai, Bin Chen, Zhi Wang, Shu-Tao Xia

ICCV 2023 · Point clouds › Point cloud classification · approach: Self-supervised point cloud pretraining · 54 citations (OpenAlex)

Read the paper

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 static VPT prompts with instance-aware dynamic prompt generation (ablation)Point-MAE prompt tuning · ScanObjectNN OBJ_BGaccuracy: 90.19 -> 92.48Improved
increase number of static VPT prompts from 10 to 156 (ablation)VPT-Deep · ScanObjectNN OBJ_BGaccuracy: 89.98 -> 90.19No clear effect
use EdgeConv graph prompt module instead of MLP or Transformer (ablation)IDPT · ScanObjectNN OBJ_BGaccuracy: 91.43 -> 92.48Improved
add IDPT instead of full fine-tuningPoint-MAE-dagger · ScanObjectNN PB_T50_RSaccuracy: 88.41 -> 88.51Improved
IDPT instead of full fine-tuning on ACTACT-dagger · ScanObjectNN PB_T50_RSaccuracy: 88.21 -> 87.65Hurt
IDPT instead of full fine-tuning on Point-MAEPoint-MAE · ScanObjectNN PB_T50_RSaccuracy: 85.18 -> 84.94Hurt
IDPT instead of full fine-tuningPoint-MAE · ModelNet40accuracy: 93.8 -> 94.4Improved
IDPT instead of static VPT on part segmentationACT · ShapeNetPartmIoU_I: 85.4 -> 85.9Improved
prompt tuning versus full fine-tuning on part segmentationACT · ShapeNetPartmIoU_I: 86.1 -> 85.9Hurt
insert dynamic prompt into deeper Transformer layers (ablation)IDPT · ScanObjectNN OBJ_BG and OBJ_ONLYaccuracyImproved
insert shared-parameter dynamic prompt in all layers (ablation)IDPT · ScanObjectNNaccuracyHurt
add IDPT prompt versus head tuning only (ablation)Point-MAE · ScanObjectNN OBJ_BGaccuracy: 87.40 -> 91.43Improved

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