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 Video

Updated: Sep 17, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

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

Published on: December 15, 2023

650

BiVM: Accurate Binarized Neural Network for Efficient Video Matting.

Haotong Qin, Xianglong Liu, Xudong Ma

    IEEE Transactions on Pattern Analysis and Machine Intelligence
    |July 2, 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

    MPQ-DMv2: Flexible Residual Mixed Precision Quantization for Low-Bit Diffusion Models With Temporal Distillation.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same author

    QuantSR+: Pushing the Limit of Quantized Image Super-Resolution Networks.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same author

    DynamicPAE: Generating Scene-Aware Physical Adversarial Examples in Real-Time.

    IEEE transactions on pattern analysis and machine intelligence·2025
    Same author

    A survey of low-bit large language models: Basics, systems, and algorithms.

    Neural networks : the official journal of the International Neural Network Society·2025
    Same author

    An empirical study of LLaMA3 quantization: from LLMs to MLLMs.

    Visual intelligence·2025
    Same author

    Construction of CdS@g-C<sub>3</sub>N<sub>4</sub> heterojunction photocatalyst for highly efficient degradation of gaseous toluene.

    The Science of the total environment·2024
    Same journal

    Relation DETR+: Exploring Explicit Position Relation Prior for Dense Prediction.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    RBF++: Quantifying and Optimizing Reasoning Boundaries across Measurable and Unmeasurable Capabilities for Chain-of-Thought Reasoning.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    CAFE: Cross-View Adaptive Fusion and Cluster Center Enhancement for Robust Multi-View Clustering.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Driving.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    Ethics-Aware Safe Reinforcement Learning for Rare-Event Risk Control in Interactive Urban Driving.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    Learning Shape Anchors for Holistic Indoor Scene Understanding.

    IEEE transactions on pattern analysis and machine intelligence·2026
    See all related articles

    This study introduces BiVM, a binarized neural network for efficient real-time video matting on edge devices. BiVM overcomes accuracy and computation limitations in existing binarized networks, enabling high-quality video matting with significantly reduced resource usage.

    Area of Science:

    • Computer Vision
    • Deep Learning
    • Neural Network Compression

    Background:

    • Deep neural networks for real-time video matting face computational limits on edge devices.
    • Binarization compresses networks but can degrade accuracy and efficiency in video matting due to encoder/decoder issues.

    Purpose of the Study:

    • To develop an accurate and resource-efficient binarized neural network for real-time video matting on edge hardware.
    • To address information degradation and redundant computation in binarized video matting networks.

    Main Methods:

    • Introduced binarized computation structures with elastic shortcuts and evolvable topologies for an improved encoder.
    • Implemented feature sparsification in the binarized decoder to focus on diverse details and reduce computation.
    • Developed a binarization-aware mimicking framework with an information-guided strategy.

    Related Experiment Videos

    Last Updated: Sep 17, 2025

    Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
    03:31

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

    Published on: December 15, 2023

    650

    Main Results:

    • BiVM significantly outperforms existing binarized video matting networks, showing a 16.67 MAD improvement on the VM dataset.
    • Achieved comparable visual quality to full-precision counterparts.
    • Demonstrated substantial savings: 14.3x in computation and 21.6x in storage costs.
    • Evaluated performance on ARM CPU, confirming suitability for resource-constrained scenarios.

    Conclusions:

    • BiVM offers an effective solution for real-time video matting on edge devices.
    • The proposed architecture and optimization strategies enable high accuracy and efficiency in binarized video matting.
    • BiVM has strong potential for practical deployment in applications like online conferences and short-form video production.