Please use this identifier to cite or link to this item:
|Title:||Development of an image recognition algorithm for VLSI circuits||Authors:||Dong, Liang||Keywords:||DRNTU::Engineering::Electrical and electronic engineering::Integrated circuits||Issue Date:||2012||Abstract:||This is the final report for final year project “Development of an Image Recognition Algorithm for VLSI circuits”. It includes the introduction, the literature review the implementation process and the result of the project of the two semesters. Nowadays, electronic manufacturing technology has been developed tremendously and it allows the creation of monolithic integrated circuits that involves millions of transistors. The complexity of integrated circuits (ICs) with multimillion gates poses great challenge to computer-aided design (CAD) tools and inevitably calls for experts from multi-disciplinary background in understanding the nature of an IC. A viable solution could be an intelligent image recognition system employed to extract and recognize sub-circuits from images of practical very large scale integrated (VLSI) circuits. In this project, MATLAB is used as the platform to implement the design. By using the functions provided in imaging processing toolbox in MATLAB, the implementation of the design can be built without using low level functions. For VLSI circuits, it has three categories of layer images, which are the substrate layer, the poly silicon layer and the metal layer. This project is mainly focused on substrate layer and poly silicon layer feature extraction. The objective is to identify features on both layers. From August 2011 to October 2011, the implementation of the MATLAB code to identify the poly line on both substrate layer and poly silicon layer is complete. From January 2012 to March 2012, image segmentation algorithm, feature extractions and classification were researched and implemented. Some features of the processed image are reviewed, such as pixel color value, segment area and so on. Some classification algorithms such as k nearest neighbor are learned for pixel classification.||URI:||http://hdl.handle.net/10356/49647||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Student Reports (FYP/IA/PA/PI)|
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.