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

Implicit hierarchical temporal-spatial residual model for long-term video prediction.

Guiqin Wang1, Peng Zhao1, Haoran Guo1

  • 1School of Computer Science and Technology, Xi'an Jiaotong University, Xi'an, Shaanxi, 710049, China; National Engineering Laboratory for Big Data Analytics, Xi'an Jiaotong University, Xi'an, Shaanxi, 710049, China.

Neural Networks : the Official Journal of the International Neural Network Society
|February 21, 2026
PubMed
Summary

Related Concept Videos

Residuals and Least-Squares Property01:11

Residuals and Least-Squares Property

9.6K
The vertical distance between the actual value of y and the estimated value of y. In other words, it measures the vertical distance between the actual data point and the predicted point on the line
If the observed data point lies above the line, the residual is positive, and the line underestimates the actual data value for y. If the observed data point lies below the line, the residual is negative, and the line overestimates the actual data value for y.
The process of fitting the best-fit...
9.6K

You might also read

Related Articles

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

Sort by
Same author

Integrative Multi-Omics Analysis Reveals Nutritional Metabolite Diversity and Regulatory Mechanisms in <i>Ocimum basilicum</i>.

Life (Basel, Switzerland)·2026
Same author

A novel ABCD1 frameshift mutation detected in a Chinese male with adrenomyeloneuropathy.

Neurological sciences : official journal of the Italian Neurological Society and of the Italian Society of Clinical Neurophysiology·2026
Same author

Discovering multiscale deep formulas in complex systems via neural-guided lambda calculus.

Nature communications·2026
Same author

The investigation on tensile failure of active waveguide structure by acoustic emission measurements.

Scientific reports·2026
Same author

Research on the Formability of 2A12 Aluminum Alloy Sheet During High-Speed Hot Gas Bulging.

Materials (Basel, Switzerland)·2026
Same author

Toward Ultrasensitive Electrochemical Detection of Ammonia Nitrogen in Drinking Water: PtCo Alloy Nanosheet on Self-Supported Carbon Cloth.

Sensors (Basel, Switzerland)·2026

Predicting future video frames is hard due to errors growing over time. This study introduces a Hierarchical Temporal-Spatial Residual Model to improve long-term video prediction by capturing residual data distributions.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Artificial Intelligence

Background:

  • Long-term video prediction is hindered by prediction ambiguities and error amplification.
  • Existing methods struggle with hierarchical spatial-temporal representations and varying spatial distributions.
  • Hierarchical spatial modeling shows promise in video spatial analysis.

Purpose of the Study:

  • To propose a novel Hierarchical Temporal-Spatial Residual Model for enhanced long-term video prediction.
  • To improve the representation of stochastic features in videos by capturing residual distributions.
  • To enhance generalization across diverse spatial distributions in video data.

Main Methods:

  • Developed a Hierarchical Temporal-Spatial Residual Model.
Keywords:
Hierarchical residual modelLong-term video predictionSpatial reconstructionStochastic feature representation

Related Experiment Videos

  • Utilized a hierarchical residual generative model to improve latent state space for spatial feature capture.
  • Explicitly modeled the residual nature of data and aligned approximate posterior with prior.

    • Main Results:

    • The proposed model significantly boosts performance in long-term video prediction tasks.
    • Achieved superior performance compared to temporal model-based and convolutional neural network-based approaches on three datasets.
    • Demonstrated improved generalization across diverse spatial distributions.

    Conclusions:

    • The Hierarchical Temporal-Spatial Residual Model effectively addresses challenges in long-term video prediction.
    • The method provides a richer representation of stochastic video features and enhances spatial generalization.
    • The publicly available code facilitates further research in implicit video prediction.