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: Mar 15, 2026

Design and Analysis for Fall Detection System Simplification
08:05

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

11.2K

An AI-Driven Multimodal Sensor Fusion Framework for Fraud Perception in Short-Video and Live-Streaming Platforms.

Ruixiang Zhao1, Xuanhao Zhang1, Jinfan Yang1,2

  • 1National School of Development, Peking University, Beijing 100871, China.

Sensors (Basel, Switzerland)
|March 14, 2026
PubMed
Summary

Related Concept Videos

Force Classification01:22

Force Classification

2.6K
Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
2.6K

You might also read

Related Articles

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

Sort by
Same author

Assessing the health utility values of patients with six common cancers in China using the QLU-C10D: A cross-sectional survey.

iScience·2026
Same author

Screening of Superior Lactic Acid Bacteria from Sourdough and Optimization of Freeze-Dried Bacterial Powder Preparation Technology for Steamed Buns.

Foods (Basel, Switzerland)·2026
Same author

A Multimodal Sensor-Based Self-Supervised Learning Framework for Low-Noise System State Prediction and Anomaly Detection.

Sensors (Basel, Switzerland)·2026
Same author

Targeting DGAT1 reprograms lipid landscape and restores CD8⁺ T cell immunity in pancreatic cancer.

Nature communications·2026
Same author

Continuous-wave organic laser enabled by monolayer molecular crystal.

Nature communications·2026
Same author

An interpretable <sup>1</sup>⁸F-PSMA PET/CT-MRI model for optimizing prostate biopsy decisions.

European journal of nuclear medicine and molecular imaging·2026
This summary is machine-generated.

This study introduces an AI framework for detecting fraud in short videos by analyzing multimodal sensor data over time. The model effectively identifies early-stage fraudulent activities, outperforming existing methods.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Data Science

Background:

  • Short-video platforms and live-streaming commerce generate complex multimodal sensor data.
  • Traditional fraud detection methods struggle with the temporal dependencies and cross-modal coupling of this data, especially in early stages.
  • Detecting subtle, evolving fraudulent signals across heterogeneous data streams remains a significant challenge.

Purpose of the Study:

  • To propose an artificial intelligence-driven multimodal sensor perception framework for temporal fraud detection in short-video environments.
  • To address the limitations of conventional unimodal or static analytical paradigms in capturing early-stage anomalous cues.
  • To enable proactive fraud warning by characterizing the progressive intensification of fraudulent activities.

Main Methods:

Keywords:
cross-modal temporal attentionmulti-sensor fusionmultimodal behavioral analyticsmultimodal perception modelingsequential signal processing

Related Experiment Videos

Last Updated: Mar 15, 2026

Design and Analysis for Fall Detection System Simplification
08:05

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

11.2K
  • Developed a multimodal temporal alignment module to synchronize sensor signals with varying granularities.
  • Constructed a shared temporal encoding network for learning evolution-aware representations from multimodal sequences.
  • Introduced a cross-modal temporal attention fusion mechanism for dynamic weighting of sensor contributions.
  • Created a fraud evolution modeling and early risk prediction module for assessing risk with incomplete temporal data.

Main Results:

  • The proposed AI framework achieved high performance metrics: 0.941 accuracy, 0.865 precision, 0.812 recall, 0.838 F1 score, and 0.956 AUC.
  • Demonstrated significant outperformance compared to text-based, vision-based, temporal, and conventional multimodal baselines.
  • Maintained performance advantages in early-stage detection (first 30% of content), with 0.812 precision, 0.704 recall, and 0.754 F1 score.

Conclusions:

  • The AI-driven sensing framework is effective for temporal fraud detection in short-video environments.
  • The model's ability to capture weak anomalous cues and evolving fraud patterns across multimodal data is validated.
  • The framework offers a robust solution for proactive fraud warning and risk assessment in dynamic online ecosystems.