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: Sep 30, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

TSGB: Target-Selective Gradient Backprop for Probing CNN Visual Saliency.

Lin Cheng, Pengfei Fang, Yanjie Liang

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society

|March 11, 2022

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

A NACHT domain-containing protein Ncp is required for appendage-associated unconventional protein secretion in the fungus <i>Penicillium herquei</i>.

iScience·2026

Same author

Global burden of bipolar disorder in 204 countries and territories, 1990-2021: a systematic analysis of temporal trends, demographic disparities, and SDI associations for the Global Burden of Disease Study 2021.

Psychiatry research·2026

Same author

A mechanosensitive ion channel is required for nematode predation in nematode-trapping fungi.

mBio·2026

Same author

Robust Model Fitting via Motion-Aware Pyramid Transformer-Guided Preference Filtering and Consensus Smoothing.

IEEE transactions on pattern analysis and machine intelligence·2026

Same author

Evolocumab protects mice from radiation-induced damage by affecting hematopoietic cell lineage.

Genomics·2026

Same author

The tango of immune and neural cells: orchestrating neuroinflammation mediated brain disorders.

Frontiers in immunology·2026

Same journal

Hyperbolic Cycle Alignment for Infrared-Visible Image Fusion.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Learning Gaze Synthesizer via 3D-eye Controlled Diffusion and Cross-domain Feature Alignment.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Underlying Semantic Diffusion for Effective and Efficient In-Context Learning.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

DiffRES: Unleashing Text-to-Image Diffusion Models for Generative Referring Expression Segmentation without Information Leakage.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Location Matters: Frequency-Spatial Dual Space Adaptation for Cross-Domain Few-Shot Segmentation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

BayeTopo: Bayesian-based Topology-guided Learning for Vascular Imaging Segmentation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

See all related articles

This study introduces Target-Selective Gradient Backprop (TSGB), a new visual saliency method for deep neural networks. TSGB generates fine-grained explanations for convolutional neural networks, improving target class selectivity.

Area of Science:

Artificial Intelligence
Computer Vision
Deep Learning

Background:

Deep neural network interpretability is crucial.
Visual saliency methods aid in understanding convolutional neural networks (CNNs).
Single backward pass methods offer speed advantages but lack target specificity.

Purpose of the Study:

To address the challenge of generating fine-grained, target-selective saliency maps using single backward pass methods.
To improve the interpretability of deep neural networks by focusing on specific target classes.

Main Methods:

Propose Target-Selective Gradient Backprop (TSGB), a novel visual saliency method.
TSGB utilizes rectification operations to enhance target class emphasis and gradient propagation.
The method comprises two components: TSGB-Conv for convolutional layers and TSGB-FC for fully-connected layers.

More Related Videos

Investigating Object Representations in the Macaque Dorsal Visual Stream Using Single-unit Recordings

Investigating Object Representations in the Macaque Dorsal Visual Stream Using Single-unit Recordings

Published on: August 1, 2018

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Published on: January 23, 2017

Related Experiment Videos

Last Updated: Sep 30, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Investigating Object Representations in the Macaque Dorsal Visual Stream Using Single-unit Recordings

Investigating Object Representations in the Macaque Dorsal Visual Stream Using Single-unit Recordings

Published on: August 1, 2018

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Published on: January 23, 2017

Main Results:

TSGB effectively generates target-selective and fine-grained saliency maps.
Experiments on ImageNet and Pascal VOC datasets demonstrate superior performance over existing methods.
The proposed method achieves more accurate and reliable visual explanations.

Conclusions:

TSGB offers a significant advancement in creating faithful saliency maps for CNNs.
The method successfully mitigates the limitations of existing single backward pass techniques.
TSGB provides a more precise and efficient approach to interpreting deep learning models.