Homepage of Jessie Burger

Distributed Mentor Project

Goals

Introduction

Project

Become familiar with ongoing Computer Vision research at Boston University.
Become familiar with and capable of implementing standard Computer Vision techniques, such as image differencing for motion detection and thresholding for conversion of grayscale images into binary images which identify objects of interest.
Aid graduate student John Magee in showing proof of concept of gaze detection for his Fast face and Eye Detection project.
Aid graduate student Stephen Crampton with his Finger Counter project.
Install and learn the Computer Vistion library OpenCV, which provides a wide array of Computer Vision algorithms such as motion detection and face detection.
Create benchmark testing suite of still images and video. Testing suite should include a variety of resolutions and activities which will enable an accurate analysis of Computer Vision Functions.
Utilize existing timing functions, either C++ code which determines time in milliseconds or assembly code which determines the number of clock cycles, to create a table of timing values for a wide array of both primitive and complex Computer Vision algorithms which can then be used for scheduling purposes.
Experiment with ways to vary parameters of algorithms in such a way as to reduce their running time, such as sampling every other pixel instead of every pixel to reduce accuracy but increase speed, and incorporate such information into the timing table.
Analyze results and pursue further research as time permits.
Write a paper for conference or journal publication.