Title : Keyhole status prediction based on voting ensemble convolutional neural networks and visualization by Grad-CAM in PAW

Hits :

Affiliation of Author(s):材料科学与工程学院

Title of Paper:Keyhole status prediction based on voting ensemble convolutional neural networks and visualization by Grad-CAM in PAW

Journal:Journal of manufacturing processes

Key Words:Convolution neural network;Keyhole plasma arc welding;Majority voting;Model visualization;Topside weld pool image

First Author:周方正

Document Code:1557660941995376641

Volume:80

Page Number:805-815

Number of Words:8000

Translation or Not:no

Date of Publication:2022-08-01