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 15, 2026

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

Esaliency (extended saliency): meaningful attention using stochastic image modeling.

Tamar Avraham¹, Michael Lindenbaum

¹Computer Science Department, Technion-Israel Institute of Technology, Haifa 32000, Israel. tammya@cs.technion.ac.il

IEEE Transactions on Pattern Analysis and Machine Intelligence

|March 13, 2010

Summary

This study introduces a novel bottom-up attention mechanism for computer vision. It uses a validated stochastic model to predict image region saliency, improving image analysis efficiency.

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

Dataset of polarimetric images of mechanically generated water surface waves coupled with surface elevation records by wave gauges linear array.

Data in brief·2025

Same author

Depth perception in autostereograms: 1/f noise is best.

Journal of the Optical Society of America. A, Optics, image science, and vision·2016

Same author

The cues in “dependent multiple cue integration for robust tracking” are independent.

IEEE transactions on pattern analysis and machine intelligence·2014

Same author

Blind adaptive sampling of images.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2011

Same author

Nonnegative Matrix Factorization with Earth Mover's Distance Metric for Image Analysis.

IEEE transactions on pattern analysis and machine intelligence·2011

Same author

Dense mirroring surface recovery from 1D homographies and sparse correspondences.

IEEE transactions on pattern analysis and machine intelligence·2011

Area of Science:

Computer Vision
Artificial Intelligence
Image Processing

Background:

Computer vision attention mechanisms allocate computational resources to important image parts, accelerating analysis.
Traditional models often attempt to mimic human attention, which can be complex and computationally intensive.

Purpose of the Study:

To propose a novel, mathematically defined bottom-up attention mechanism for computer vision.
To develop a stochastic model for estimating image region saliency without relying on human attention mimicry.
To improve the efficiency and effectiveness of image analysis processes.

Main Methods:

A validated stochastic model was developed to estimate the probability of image parts being of interest (saliency).
The model incorporates observations like visual similarity and the presence of few salient objects (relaxed global exceptions).

Related Experiment Videos

Last Updated: Jun 15, 2026

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

A graphical model approximation was used following preattentive segmentation to identify salient image segments.

Main Results:

The proposed saliency model effectively quantifies intuitive observations about image content.
Experiments on natural scenes demonstrated the method's advantages over existing approaches.
The bottom-up attention mechanism showed improved performance in identifying regions of interest.

Conclusions:

The new stochastic saliency model offers a mathematically rigorous and efficient approach to computer vision attention.
This method enhances image analysis by focusing computational resources on relevant image regions.
The proposed bottom-up attention mechanism presents a promising alternative to traditional, human-mimicking attention models.