Wayne, Lin Wei-Cheng

Current Status

MBA Candidate, Class of 2009, UCLA Anderson School of Management

Research Interest

Digital Watermark

Image/Video Segmentation

Mathematical Morphology

Education:

• MS, in Communication and Multimedia Lab of CSD in NTU
• BS, Class of 1994, CSIE, National Taiwan University
• Chien-Kuo Senior High School
• Da-Ji Junior High School
• Da-Ji Elementary School

Experience:

• Partner of HoEArt Creative Corporation
• Contract Research Assistant in Communication and Multimedia Lab, NTU
• Advanced Researcher, Advanced Technology Center of Arcadyan Technology Corporation
• Researcher in Philips Research East Asia - Taipei
• Part-time Research Engineer of PREA-T
• Part-time Programmer of EeRise Tech.

Previous Projects :

• Fingerprint Recognition (1998) (DOWNLOAD PDF IN CHINESE) -- 7th NTU Engineering Technology Contest with the second prize.

      A system for fingerprint recognition based on the minutiae of the ridges is introduced in this paper.   The system is divided into two parts, minutiae extracion and minutiae matching.   In the first part, the ridges of the fingerprints are extracted to find the end points and bifurcations of the ridges.   In the second part, an alignment-based matching algorithm is applied to figure out the similarity between two fingerprints.   Each stage of the system for the rigde-based fingerprint recognition is discussed in detail and implemented by choosing a suitable algorithm.   Some operations of mathematical morphology are also introduced into the first part of the system to enhance the accuracy of minutiae extraction.

• HTML Classification (1998)

      The goal of this project is to re-consider the learning methods introduced in the AI course and use the 4 University data set as the training data to experiment on the classification of html documents.   While making experiments in the project, two classifiers have been implemented : hybrid html classification of term-probability and document-probability, and term-weight-based html classification.

• Mathematical Morphology and Its Applications on Image Segmentation (1998-2000) (DOWNLOAD PDF) -- Master Thesis

      Mathematical morphology provides a systematic approach to analyze the geometric characteristics of signals or images, and has been applied widely to many applications such as edge detection, object segmentation, noise suppression. In this paper, a morphology-based supervised segmentation system is proposed. To know where a semantic object resides, the user could click the mouse near the boundary to indicate its rough definition. The proposed system would first locate the searching area, and then the units in it are classified into "object" and "non-object" to find out the continuous contours by means of a multi-valued watershed algorithm using a hierarchical queue. An adaptive morphological operator based on edge strength, which is computed by a multiscale morphological gradient algorithm, is proposed to tolerate the error of user assistance such that the searching area is created correctly. Moreover, besides using pixels as the basic units for classification, the homogeneous regions produced by the morphological watershed transformation can also be used to further reduce the time of classification. Due to the capability of the adaptive operator, the proposed system could be extended to video object tracking without motion information from motion estimator, if the moving of the object does not exceed the searching bound of the adaptive operator.

Publications:

• Wei-Cheng Lin, Mathematical Morphology and Its Applications on Image Segmentation. Master Thesis, CSIE Dept. of NTU, June 2000.

• Wayne Lin, Chi-Hui Huang and Ja-Ling Wu, "A Novel Supervised Image/Video Segmentation System Based on Mathematical Morphology", Proceedings of 2001 IEEE-EURASIP Workshop on Nonlinear Signal and Image Processing , Baltimore, USA, June 2001.