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

Zero-shot incremental learning using spatial-frequency feature representations.

Jie Ren1, Yang Zhao1, Weichuan Zhang2

  • 1Xi'an Polytechnic University, Xi'an, Shaanxi, China.

Scientific Reports
|March 29, 2025
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

Pan-cancer Distant Metastasis Prediction Based on Graph Neural Network.

Interdisciplinary sciences, computational life sciences·2026
Same author

Gene-level gut microbiome signatures as predictive biomarkers for response to immune checkpoint inhibitors across multiple cancer types.

Gut microbes·2026
Same author

PGST: A prototype-guided parameter-efficient network for spatial transcriptomics prediction.

IEEE journal of biomedical and health informatics·2026
Same author

BiChemoCLAM: a weakly supervised multimodal framework for chemotherapy response prediction.

Briefings in bioinformatics·2026
Same author

Printed Circuit Board Defect Detection Based on Lightweight Deep Learning Fusion Model.

Sensors (Basel, Switzerland)·2025
Same author

TCN-RDP: Predicting Drug-Induced Liver Injury from Time-Series Toxicogenomic Data.

Journal of chemical information and modeling·2025
Same journal

Turbulent flow in a vortex separator with a directed pipe inlet.

Scientific reports·2026
Same journal

Systematic characteristic evaluation of clay-based cementitious material derived from calcium carbide residue and waste tile powder.

Scientific reports·2026
Same journal

Retraction Note: Improvement of a rapid diagnostic application of monoclonal antibodies against avian influenza H7 subtype virus using Europium nanoparticles.

Scientific reports·2026
Same journal

Applying large language models to spam detection in the Kazakh low-resource language setting.

Scientific reports·2026
Same journal

An open-source 3D printing system enabling in-situ freeze-thaw processing of hydrogels.

Scientific reports·2026
Same journal

An enhanced EfficientNet framework for automated waste classification using cosine annealing and label smoothing.

Scientific reports·2026
See all related articles

This study introduces a new network for zero-shot incremental learning, significantly reducing data forgetting. The spatial-frequency feature representation network (SFFRNet) enhances generalization to new classes by capturing crucial image domain information.

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Machine Learning

Background:

  • Zero-shot incremental learning aims for models to learn new classes without forgetting old ones.
  • Catastrophic forgetting is a major challenge due to the semantic gap between old and new classes.
  • Existing methods struggle to capture essential information from diverse image sample domains.

Purpose of the Study:

  • To propose a novel Spatial-Frequency Feature Representation Network (SFFRNet) for class incremental learning.
  • To address catastrophic forgetting and improve generalization to new classes in zero-shot learning scenarios.
  • To enhance the extraction of significant information from sample image domains.

Main Methods:

  • Developed SFFRNet with integrated Spatial Feature Extraction (SFE) and Frequency Feature Extraction (FFE) modules.
Keywords:
Attention mechanismDiscrete cosine transformIncremental learningZero-shot learning

Related Experiment Videos

  • Implemented a novel approach to extract combined spatial-frequency feature representations from images.
  • Evaluated the model's performance on benchmark datasets like CUB 200-2011 and CIFAR-100.

    • Main Results:

    • SFFRNet effectively extracts spatial-frequency features, improving image classification accuracy.
    • The proposed method significantly alleviates catastrophic forgetting in incremental learning.
    • Experimental results show superior performance compared to existing state-of-the-art incremental learning algorithms.

    Conclusions:

    • SFFRNet offers a robust solution for zero-shot class incremental learning.
    • The spatial-frequency feature representation is key to overcoming catastrophic forgetting.
    • The approach demonstrates strong potential for real-world applications requiring continuous learning.