1 / 8

Adam Yeh

Adam Yeh. UCF Computer Vision REU Week 1. About Me. Cornell University, Computer Science, c/o 2009 Coursework: Data Structures/Algorithms Functional Programming Java, C, Matlab Computation, Information, and Intelligence (TA also) Math courses too: Discrete Math Number Theory

Download Presentation

Adam Yeh

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Adam Yeh UCF Computer Vision REU Week 1

  2. About Me • Cornell University, Computer Science, c/o 2009 • Coursework: • Data Structures/Algorithms • Functional Programming • Java, C, Matlab • Computation, Information, and Intelligence (TA also) • Math courses too: • Discrete Math • Number Theory • Group Theory • Multivariable/DiffEq/Linear Algebra • Goals: • Grad school-Ph.D? Math/CS • Research career… • This is why I’m here…

  3. This Week • Ran Canny/Sobel C code • Ran FingerFinder code (previous REU’s work) • Started writing Canny code in Matlab • For practice in coding (haven’t done it in awhile) • To understand algorithm better • Currently unfinished (still need double thresholding) • Tried an optimization of Java Canny code • For double thresholding, use BFS to find connected edges instead of repeatedly scanning entire image • Only slightly faster • Algorithmically should be faster, but more overhead • Background Research

  4. 1024*1280 pixel image, Threshold=1, 15% • Canny runs in ~9.7s • Modified Canny runs in ~9.5s • Eventually runs out of memory…

  5. Canny: Changing Sigma • Sigmas of 1, 2, 3 (15%) • More blending means less defined edges aren’t caught

  6. Canny: Changing Threshold % • Threshold determined by percentile (best edges are top 15, 5, and 35%) • Bottom threshold is .35 of top threshold • High threshold means only defined edges caught • Low threshold results in noise

  7. Possible Topics • Initially: lip reading as improvement to speech recognition • Rejected because of emphasis on sound as well • Facial Recognition • Work has been started by Dr. Lobo’s students • Modify finger finder? • Learning algorithms: AdaBoost+Support Vector Machines? • 7 basic expressions: Neutral, anger, disgust, fear, joy, sadness, and surprise

  8. Next Week • Background information: • Current Facial Recognition research? • More machine learning/pattern matching techniques • Modify finger code for specific features • Different parts of face e.g. eyes, nose, mouth • Try OpenCV implementation of Viola/Jones • Find database of test images

More Related