中美计算机科学研究中心特邀美国新泽西理工学院计算机教授Frank Shih做学术报告，欢迎全校师生参加。
Title:
Developmentof Deep Learning Framework for Image Morphology andApplications to Ecology and Conservation
Speaker:
Dr. Frank Shih, Professor of Computer Science, New Jersey Institute of Technology, USA
Abstract:
The importance and impactof mathematical morphology have been well documented in the areas ranging fromautomated vision detection and inspection to object recognition, imageanalysis, and pattern recognition. In this talk, we will present a noveldeep-learning framework for automatically learning applicable morphologicaloperations with suitably weighted structuring elements in a massive image datasetting. To verify the correctness of the deep learning approximation, weconstruct a MorphNet to learn the targets of dilated or eroded images and the networkweights correctly. There has been a rapid development of automated speciesidentification systems based on imagery. We apply deep learning framework to classify19 classes of insect-wing images using different deep learning models. Twodatabases of the wings ofvarious native bee species imagery are used to discover optimal imagepreprocessing steps by deep learning. Since the datasets contain small amountsof images, we enlarge the datasets into a large scale by applying imageaugmentation and generalization techniques. Through extensive training, we areable to achieve a very high testing accuracy rate of 98.34%.
时间：4月12日下午 15：30-16：30
地点：中美计算机科学研究中心报告厅
报告人简介
ProfessorFrank Shih received M.S. from State University of New York, Stony Brook,U.S.A., in 1984, and Ph.D. from Purdue University, West Lafayette, IN, U.S.A.,in 1987. He is presently a professor jointly appointed in the Department ofComputer Science, the Department of Electrical and Computer Engineering, andthe Department of Biomedical Engineering at New Jersey Institute of Technology,Newark, New Jersey. He is the Director of Computer Vision Laboratory. He is aninternationally renowned scholar and serves on the Editorial Board of 12international journals. He served as a steering member, committee member, andsession chair for numerous professional conferences and workshops.
Dr.Shih has made significant contributions to mathematical morphology, patternrecognition, and information hiding, focusing on the security and robustness ofdigital watermarking and steganography. He has authored four books: “DigitalWatermarking and Steganography,” “Image Processing and MathematicalMorphology,” “Image Processing and Pattern Recognition,” and “MultimediaSecurity: Watermarking, Steganography, and Forensics.” He has published over 135journal papers, 100 conference papers, and 20 book chapters. His researchinterests include image processing, computer vision, watermarking andsteganography, digital forensics, sensor networks, pattern recognition,bioinformatics, information security, robotics, fuzzy logic, and neuralnetworks.
计算机与软件学院
中美计算机科学研究中心
