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 Experiment Videos

Manifold regularized multitask learning for semi-supervised multilabel image classification.

Yong Luo1, Dacheng Tao, Bo Geng

  • 1Key Laboratory of Machine Perception (Ministry of Education), School of Electronics Engineering and Computer Science, Peking University, Beijing 100871, China. yluo180@gmail.com

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
|September 22, 2012
PubMed
Summary
This summary is machine-generated.

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

A built-in electric field induces a high-performance hydrogen evolution reaction on a self-supporting MoO<sub>2</sub>-NiP/NF heterojunction.

Nanoscale·2026
Same author

A few-shot high-resolution remote sensing image semantic segmentation method.

Scientific reports·2026
Same author

In-silico study of machine learning discovers de novo lipogenesis for predicting prognosis and immunotherapy responses in kidney renal clear cell carcinoma.

Biology direct·2026
Same author

Correction: Correction to: Tumor-associated macrophage-derived exosomes transmitting miR-193a-5p promote the progression of renal cell carcinoma via TIMP2-dependent vasculogenic mimicry.

Cell death & disease·2025
Same author

Synergistically promoting proton-coupled electron transfer of oxygen reduction with dual atomic sites on high-curvature carbon onions for highly efficient Zn-air batteries.

Nanoscale·2025
Same author

Upregulated NNMT sustains angiogenesis in clear cell renal cell carcinoma through activation of the SDF-1/CXCR4-ERK-CCL20 axis.

Cellular oncology (Dordrecht, Netherlands)·2025
Same journal

Mask-guided Asymmetric Contrastive and Semantic Alignment for Unsupervised Person Re-Identification.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

Hyperbolic Cycle Alignment for Infrared-Visible Image Fusion.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

Learning Gaze Synthesizer via 3D-eye Controlled Diffusion and Cross-domain Feature Alignment.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

Underlying Semantic Diffusion for Effective and Efficient In-Context Learning.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

DiffRES: Unleashing Text-to-Image Diffusion Models for Generative Referring Expression Segmentation without Information Leakage.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

Location Matters: Frequency-Spatial Dual Space Adaptation for Cross-Domain Few-Shot Segmentation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
See all related articles

Classifying images with multiple labels using limited data is difficult. A new manifold regularized multitask learning (MRMTL) algorithm effectively addresses this challenge by learning a shared subspace for multiple tasks, improving image classification accuracy.

Area of Science:

  • Computer Science
  • Machine Learning
  • Artificial Intelligence

Background:

  • Multi-label image classification with limited labeled data presents a significant challenge.
  • Traditional methods using individual binary classifiers with manifold regularization struggle with high-dimensional visual features.
  • Existing manifold regularization techniques are insufficient for controlling model complexity in complex image datasets.

Purpose of the Study:

  • To propose a novel manifold regularized multitask learning (MRMTL) algorithm for improved multi-label image classification.
  • To address the limitations of existing methods in handling high-dimensional data and limited samples.
  • To enhance the control of model complexity in multi-label classification tasks.

Main Methods:

  • Developed a manifold regularized multitask learning (MRMTL) algorithm.

Related Experiment Videos

  • MRMTL learns a shared discriminative subspace across multiple classification tasks.
  • Incorporated manifold regularization to ensure smoothness along the data manifold within the shared hypothesis space.

    • Main Results:

    • MRMTL demonstrated effectiveness in multi-label image classification tasks.
    • Experiments were conducted on the PASCAL VOC'07 (20 classes) and MIR (38 classes) datasets.
    • The proposed MRMTL algorithm outperformed popular existing image classification algorithms.

    Conclusions:

    • MRMTL effectively controls model complexity by leveraging the common structure of multiple tasks.
    • The algorithm enhances classification performance by learning a shared subspace and utilizing manifold regularization.
    • MRMTL offers a promising solution for challenging multi-label image classification scenarios with limited labeled data.