Grad++ScoreCAM: Enhancing Visual Explanations of Deep Convolutional Networks Using Incremented Gradient and Score- Weighted Methods

We propose a novel method that combines the strengths of two popular class activation mapping techniques, GradCAM++ and ScoreCAM, to improve the interpretability and localization of convolutional neural networks (CNNs). Our proposed method, called “Grad++-ScoreCAM”, first utilizes the...

Cijeli opis

Bibliografski detalji

Glavni autori:	Shafiullah Soomro, Asim Niaz, Kwang Nam Choi
Format:	Artigo
Jezik:	Inglês
Izdano:	IEEE 2024-01-01
Serija:	IEEE Access
Teme:	Class activation convolutional neural networks decision-making recognition decision interpretation
Online pristup:	https://ieeexplore.ieee.org/document/10506957/
Oznake:	Dodaj oznaku Bez oznaka, Budi prvi tko označuje ovaj zapis!