Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

RAD51 gene is associated with advanced age-related macular degeneration in Chinese population.

Clinical biochemistry·2013
Same author

Immunization against recombinant GnRH-I alters ultrastructure of gonadotropin cell in an experimental boar model.

Reproductive biology and endocrinology : RB&E·2013
Same author

Multi-class constrained normalized cut with hard, soft, unary and pairwise priors and its applications to object segmentation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2013
Same author

Comparison of genomic and amino acid sequences of eight Japanese encephalitis virus isolates from bats.

Archives of virology·2013
Same author

Regulation of dendritic cell differentiation in bone marrow during emergency myelopoiesis.

Journal of immunology (Baltimore, Md. : 1950)·2013
Same author

Separation of mandelic acid and its derivatives with new immobilized cellulose chiral stationary phase.

Journal of Zhejiang University. Science. B·2013

Related Experiment Video

Updated: Jul 5, 2025

Measuring the Structure, Composition, and Change of Underwater Environments with Large-area Imaging
09:19

Measuring the Structure, Composition, and Change of Underwater Environments with Large-area Imaging

Published on: April 18, 2025

485

Point-to-Pixel Prompting for Point Cloud Analysis With Pre-Trained Image Models.

Ziyi Wang, Yongming Rao, Xumin Yu

    IEEE Transactions on Pattern Analysis and Machine Intelligence
    |January 16, 2024
    PubMed
    Summary

    This study introduces Point-to-Pixel Prompting and Pixel-to-Point distillation to transfer knowledge from 2D to 3D vision. This method significantly improves 3D point cloud analysis performance, achieving state-of-the-art results.

    More Related Videos

    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging
    11:38

    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging

    Published on: October 4, 2024

    577
    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    2.8K

    Related Experiment Videos

    Last Updated: Jul 5, 2025

    Measuring the Structure, Composition, and Change of Underwater Environments with Large-area Imaging
    09:19

    Measuring the Structure, Composition, and Change of Underwater Environments with Large-area Imaging

    Published on: April 18, 2025

    485
    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging
    11:38

    Author Spotlight: Enhancing PSC-to-Functional Cell Differentiation Using ML Models Based on Live-Cell Bright-Field Imaging

    Published on: October 4, 2024

    577
    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    2.8K

    Area of Science:

    • Computer Vision
    • Machine Learning
    • 3D Data Analysis

    Background:

    • Pre-training large models is successful in 2D but underdeveloped in 3D vision.
    • Existing 3D vision methods lack efficient knowledge transfer from 2D domains.

    Purpose of the Study:

    • To develop a novel method for transferring pre-trained 2D knowledge to 3D point cloud analysis.
    • To enhance the performance and efficiency of 3D vision models.

    Main Methods:

    • Point-to-Pixel Prompting: Transforms point clouds into images preserving geometry and color.
    • Pixel-to-Point Distillation: Uses pre-trained 2D models as teachers to distill knowledge into 3D models.
    • Leverages shared knowledge between 2D images and 3D point clouds of the same scene.

    Main Results:

    • Achieved 90.3% accuracy in object classification on ScanObjectNN, surpassing prior work.
    • Demonstrated superior performance in scene-level semantic segmentation compared to traditional 3D methods.
    • Showcased significant improvements in inference efficiency and model capacity for point cloud analysis.

    Conclusions:

    • The proposed Point-to-Pixel Prompting and Pixel-to-Point distillation effectively transfer 2D knowledge to 3D vision tasks.
    • The method offers a scalable approach for advancing 3D point cloud understanding and analysis.