The method of performance advancement using modified neural network for test algorithm of semiconductor packages

Chang-Hyun Kim; Hong-Yeon Yu; Sung-Hoon Hong

doi:10.1117/12.784191

9 January 2008 The method of performance advancement using modified neural network for test algorithm of semiconductor packages

Chang-Hyun Kim, Hong-Yeon Yu, Sung-Hoon Hong

Author Affiliations +

Proceedings Volume 6794, ICMIT 2007: Mechatronics, MEMS, and Smart Materials; 679455 (2008) https://doi.org/10.1117/12.784191
Event: ICMIT 2007: Mechatronics, MEMS, and Smart Materials, 2007, Gifu, Japan

Abstract

The classification of defects in semiconductor packages was performed by the pattern recognition technology with modified neural network is based on image processing. The pattern recognition algorithm is composed of image processing and modified backpropagation neural network. Image processing is preprocessing method for dimensionality reduction that is input data of backpropagation neural network. And image processing is simply made of image equalization and binary image conversion and edge detection for reducing operation time. And most of algorithm of backpropagation neural network is generally used uniform train weight, but the algorithm in this research is applied to variously subdivided train weights of backpropagation neural network based on types of semiconductor packages according to kinds of defects. Through above processes, we obtained advanced result of pattern recognition about defects in semiconductor packages.

Citation Download Citation

Chang-Hyun Kim, Hong-Yeon Yu, and Sung-Hoon Hong "The method of performance advancement using modified neural network for test algorithm of semiconductor packages", Proc. SPIE 6794, ICMIT 2007: Mechatronics, MEMS, and Smart Materials, 679455 (9 January 2008); https://doi.org/10.1117/12.784191

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available