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: May 23, 2026

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

Learning where to attend with deep architectures for image tracking.

Misha Denil¹, Loris Bazzani, Hugo Larochelle

¹University of British Columbia, Vancouver, BC V6G 1Z4, Canada. mdenil@cs.ubc.ca

Neural Computation

|April 19, 2012

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

Optimisation of quantitative brain diffusion-relaxation MRI acquisition protocols with physics-informed machine learning.

Medical image analysis·2024

Same author

Competition-level code generation with AlphaCode.

Science (New York, N.Y.)·2022

Same author

Restoring and attributing ancient texts using deep neural networks.

Nature·2022

Same author

Machine behaviour.

Nature·2019

Same author

Distinct ECG Phenotypes Identified in Hypertrophic Cardiomyopathy Using Machine Learning Associate With Arrhythmic Risk Markers.

Frontiers in physiology·2018

Same author

Building machines that learn and think for themselves.

The Behavioral and brain sciences·2018

This study introduces an attentional model for object tracking and recognition using gaze data. The model improves performance with partial information by using a Gaussian process for gaze selection in a continuous action space.

Area of Science:

Computer Vision
Machine Learning
Neuroscience

Background:

Object tracking and recognition are crucial in computer vision.
Existing models often struggle with partial information and discrete action spaces.
Neuroscience theories suggest distinct pathways for object identity and spatial control.

Purpose of the Study:

To develop an attentional model for simultaneous object tracking and recognition driven by gaze data.
To address limitations of previous models in handling partial information and continuous action spaces.
To integrate insights from neuroscience into a computational model.

Main Methods:

A dual-pathway model (identity and control) inspired by neuroscience.
Deep (factored)-restricted Boltzmann machines for object classification.

More Related Videos

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Related Experiment Videos

Last Updated: May 23, 2026

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Particle filtering for estimating object state (location, orientation, scale, speed).

Gaussian process modeling for gaze selection in a continuous action space.

Main Results:

The proposed model effectively handles foveated images with decaying peripheral resolution.
A novel gaze selection strategy using Gaussian processes significantly improves performance with partial information.
The model expands gaze selection from discrete points to a continuous domain, minimizing tracking uncertainty.

Conclusions:

The attentional model provides a robust framework for simultaneous object tracking and recognition.
Gaussian process-based gaze selection is a viable strategy for handling partial information and continuous action spaces.
The model's architecture, inspired by neuroscience, offers a promising direction for advanced perception systems.