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: Jun 6, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

Bilayer segmentation of webcam videos using tree-based classifiers.

Pei Yin¹, Antonio Criminisi, John Winn

¹Microsoft Corp., One Microsoft Way, Redmond, WA 98052, USA. peiyin@microsoft.com

IEEE Transactions on Pattern Analysis and Machine Intelligence

|November 20, 2010

Summary

This summary is machine-generated.

Related Concept Videos

Force Classification

Force Classification

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,...

You might also read

Related Articles

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

Sort by

Same author

[Development of an intensified fed-batch process to enhance recombinant antibody productivity of CHO cells based on cell cycle regulation].

Sheng wu gong cheng xue bao = Chinese journal of biotechnology·2026

Same author

Application of <sup>131</sup>I-MIBG imaging in assessing renal sympathetic overdrive in heart failure with preserved ejection fraction (HFpEF).

American journal of translational research·2026

Same author

Broadband Dielectric Response of Group-II Metal Oxide Monolayers: From Ionic to Electronic Polarization.

Micromachines·2026

Same author

Early Transient Hypothyroidism Is Associated With Long-Term Outcomes of 131I Treatment in Graves' Disease: A Retrospective Cohort Study.

Clinical endocrinology·2026

Same author

Artificial intelligence in the diagnosis of thyroid diseases: applications and challenges.

Frontiers in radiology·2026

Same author

Association of Time-to-Peak of Renal Blood Flow with Outcomes in Heart Failure: An Exploratory Comparative Analysis with Conventional Predictors.

Balkan medical journal·2026

Same journal

A Comprehensive Survey on Multimodal Recommender Systems: Taxonomy, Evaluation, and Future Directions.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

Benchmarking the Robustness of Autonomous Driving to Environmental Illusions: A Lane Perception Perspective.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

Learning Topology-Aware Representations via Test-Time Adaptation for Anomaly Segmentation.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

TraGraph-GS: Trajectory Graph-based Gaussian Splatting for Arbitrary Large-Scale Scene Rendering.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

SWIFT: A Small-World Interaction Framework for Flow-Aware Trajectory Prediction in Autonomous Driving.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

HardFlow: Hard-Constrained Sampling for Flow-Matching Models Via Trajectory Optimization.

IEEE transactions on pattern analysis and machine intelligence·2026

See all related articles

This study introduces an automatic video segmentation algorithm using a monocular webcam, achieving depth segmentation comparable to stereo cameras. The novel "motons" representation and spatial motion context improve foreground-background segmentation accuracy.

Area of Science:

Computer Vision
Machine Learning
Image Processing

Background:

Accurate foreground-background segmentation is crucial for video analysis.
Monocular depth segmentation typically requires complex initialization or stereo vision.

Purpose of the Study:

To develop an automatic, initialization-free video segmentation algorithm using a monocular webcam.
To achieve depth segmentation performance comparable to stereo camera systems.

Main Methods:

Introduced a novel motion representation called "motons."
Estimated segmentation likelihood from spatial context of motion using random forests.
Fused visual cues (motion, color, contrast, spatial priors) with a conditional random field (CRF) model.
Achieved segmentation via binary min-cut.

Related Experiment Videos

Last Updated: Jun 6, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

Main Results:

The algorithm effectively segments foreground and background layers in video frames.
Achieved accurate segmentation even with significant background motion and a stationary foreground.
Segmentation results are comparable to those from stereo vision systems.
Demonstrated effectiveness across various scenes in a videochat application.

Conclusions:

The proposed monocular webcam algorithm offers a viable alternative to stereo systems for depth segmentation.
The novel "motons" representation and spatial motion context enhance segmentation accuracy.
The algorithm is robust, requires no initialization, and performs well in diverse scenarios.