Wafer map defect pattern classification based on convolutional neural network features and error-correcting output codes

Cheng Jin

doi:10.1007/s10845-020-01540-x

Wafer map defect pattern classification based on convolutional neural network features and error-correcting output codes

2020-02-04

42

PRE

AI

DOI：10.1007/s10845-020-01540-x

Original

Original request for help

Save

Abstract

En 中文

Defect clusters on the wafer map can provide important clue to identify the process failures so that it is important to accurately classify the defect patterns into corresponding pattern types. In this research, we present an image-based wafer map defect pattern classification method. The presented method consists of two main steps: without any specific preprocessing, high-level features are extracted from convolutional neural network and then the extracted features are fed to combination of error-correcting output codes and support vector machines for wafer map defect pattern classification. To the best of our knowledge, no prior work has applied the presented method for wafer map defect pattern classification. Experimental results tested on 20,000 wafer maps show the superiority of presented method and the overall classification accuracy is up to 98.43%.

Keywords:

Wafer map

Defect pattern classification

Deep learning

Convolutional neural network

Error-correcting output codes

Support vector machine

Multi-class classification

Wafer map defect pattern classification based on convolutional neural network features and error-correcting output codes

Abstract

Journal

Organization