Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Sep 30, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Self-Training for Class-Incremental Semantic Segmentation.

Lu Yu, Xialei Liu, Joost van de Weijer

IEEE Transactions on Neural Networks and Learning Systems

|March 17, 2022

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

Trust Your Good Friends: Source-Free Domain Adaptation by Reciprocal Neighborhood Clustering.

IEEE transactions on pattern analysis and machine intelligence·2023

Same author

Generative Multi-Label Zero-Shot Learning.

IEEE transactions on pattern analysis and machine intelligence·2023

Same author

Brain responses to negated and affirmative meanings in the auditory modality.

Frontiers in human neuroscience·2023

Same author

Class-Incremental Learning: Survey and Performance Evaluation on Image Classification.

IEEE transactions on pattern analysis and machine intelligence·2022

Same author

Distributed Learning and Inference With Compressed Images.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2021

Same author

Predictors of narrative text quality in students with hearing loss.

Logopedics, phoniatrics, vocology·2021

Same journal

Intervention Feasible Region and Driver Risk Capacity Aware Human-Machine Collaborative Safe Trajectory Planning.

IEEE transactions on neural networks and learning systems·2026

Same journal

A Unified Differential Denoising Learning Framework With a Pre-Trained Model and Fuzzy Graph Networks for Drug-Drug Interaction Prediction.

IEEE transactions on neural networks and learning systems·2026

Same journal

Self-Supervised Continuous Dynamic Graph Representation Learning via Hawkes Processes.

IEEE transactions on neural networks and learning systems·2026

Same journal

cPU: Consistent Risk Estimator for Positive-Unlabeled Learning.

IEEE transactions on neural networks and learning systems·2026

Same journal

Tuning-Free Latent Diffusion Models for Ultrahigh-Resolution Image Editing.

IEEE transactions on neural networks and learning systems·2026

Same journal

Hidden Data Recovery and Forecasting via Next-Generation Reservoir Computing With Multiscale Delay Selection.

IEEE transactions on neural networks and learning systems·2026

See all related articles

This study introduces a novel self-training method to combat catastrophic forgetting in class-incremental semantic segmentation using unlabeled data for knowledge rehearsal. The approach significantly improves performance on benchmark datasets.

Area of Science:

Computer Vision
Machine Learning
Deep Learning

Background:

Deep neural networks in class-incremental semantic segmentation forget previous knowledge due to lack of access to prior labeled data.
Catastrophic forgetting is a major challenge when incrementally learning new classes.

Purpose of the Study:

To propose a self-training approach using unlabeled data to mitigate catastrophic forgetting in class-incremental semantic segmentation.
To enhance knowledge rehearsal by leveraging both old and newly trained models.

Main Methods:

A temporary model is trained for the current task.
Pseudo-labels for unlabeled data are generated by fusing information from old and temporary models.
Conflict reduction and self-entropy maximization are employed to refine pseudo-labels and predictions.

More Related Videos

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

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

Published on: November 30, 2022

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Published on: March 3, 2023

Related Experiment Videos

Last Updated: Sep 30, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

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

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

Published on: November 30, 2022

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Published on: March 3, 2023

Main Results:

Achieved significant performance gains, including up to 114% relative gain on Pascal-VOC 2012 and 8.5% on ADE20K.
Demonstrated that diverse, general-purpose auxiliary data can yield substantial performance improvements.
Outperformed previous state-of-the-art methods in class-incremental semantic segmentation.

Conclusions:

The proposed self-training method effectively addresses catastrophic forgetting in class-incremental semantic segmentation.
Leveraging unlabeled data and employing conflict reduction strategies are key to successful knowledge rehearsal.
The approach shows robustness and adaptability with diverse auxiliary data.