Instructor: Dan Maynes-Aminzade (monzy at stanford dot edu), Gates Bldg. Room 386

Office Hours: Wednesdays 2:00pm - 4:00pm, or by appointment

Course Assistant: Crystal Fong (cfongc at stanford dot edu), Gates Bldg. Room B26A

Office Hours: Mondays 10:00am - 12:00pm

When emailing about the course, please include “CS377S” in the subject.

Course Newsgroup: su.class.cs377s

Please use the newsgroup for discussion and general questions about assignments.

Course Mailing List: cs377s-win0708-staff at lists dot stanford dot edu

Looking for project ideas? Browse these examples of cool computer vision projects.

Technology Demos and Research Projects

• Pittsburgh Pattern Recognition: Face Detection Demo
  http://demo.pittpatt.com/
• Automatic Sign Detection Algorithm
  http://video.google.co.uk/videoplay?docid=-3363761930395266419
• The Shrug Detector
  http://www.newscientist.com/blog/technology/2006/04/shrug-detector.html
• Peekaboom, the fun collaborative online game where you help train computer vision recognition algorithms
  http://www.peekaboom.org/
• Nouse: Use your nose as a mouse
  http://synapse.vit.iit.nrc.ca/Nouse/index2.html
• The Interactive Virtual Aerobics Trainer uses motion history images to provide feedback from a virtual instructor
  http://www.cse.ohiostate.edu/~jwdavis/CVL/Research/VirtualAerobics/aerobics.html
• A robotic hand that catches objects with quick reflexes
  http://www.k2.t.u-tokyo.ac.jp/fusion/VisualFeedbackGrasping/index-e.html
• Suveillance System Spots Violent Behavior
  http://www.newscientisttech.com/article/dn10387-surveillance-system-spots-violent-behaviour.html
• Fish-sorting Camera
  http://www.planetark.com/dailynewsstory.cfm/newsid/23814/story.htm
• Automatic Photo Pop-Up
  http://www.cs.cmu.edu/~dhoiem/projects/popup/index.html
• The DARPA Grand Challenge
  http://www.grandchallenge.org/
• Software for detecting art fakes uses color and brushstroke analysis to reveal which paintings are authentic
  http://www.newscientisttech.com/article/dn10298-detecting-art-fakes-at-a-stroke.html
• Smart Browsing of Large Art Collections using computer vision and Natural Language Processing
  http://www.coe.berkeley.edu/labnotes/1002/forsyth.html
• Automatic sports highlights: software that churns through sports video and automatically grabs the highlights
  http://www.newscientist.com/article.ns?id=dn6097
• Vision Technology for the Totally Blind: See with your Ears!
  http://www.seeingwithsound.com/javoice.htm
• The DyPERS system associates AV sequences with physical objects, so that vision-based object detection can trigger the playback of related clips
  http://vismod.media.mit.edu/vismod/demos/dypers/

Commercial Products and Computer Vision Companies

• GestureTek: Wall and table displays using simple vision-based gesture control
  http://www.gesturetek.com/
• Reactrix: company makes interactive floor ads using vision tracking
  http://reactrix.com/
• GestPoint: 3D hand gesture tracking using two cameras
  http://www.gesturetek.com/gestpoint/introduction.php
• Checker, a commerical tool for product inspection on assembly lines
  http://www.cognex.com/products/Checker/default.asp
• Semacode: Image recognition on mobile camera phones
  http://www.csclub.uwaterloo.ca/media/Semacode%20-
  %20Image%20recognition%20on%20mobile%20camera%20phones.html
• Evolution Robotics ERSP Robotic Development Platform, with downloadable demo
  http://www.evolution.com/products/ersp/
• MyHeritage Face Recognition Software
  http://www.myheritage.com/FP/Company/face-recognition.php
  http://www.saynotocrack.com/index.php/2006/10/23/celebrity-look-a-likes-part-i/
• Tyzx Stereo Range Cameras
  http://www.tyzx.com/
• ASIMO: posture, gesture, and face recognition for human-robot interaction
  http://world.honda.com/ASIMO/technology/intelligence.html
• VisionSphere: Face Recognition for biometrics and law enforcement
  http://www.visionspheretech.com/menu.htm
• Cinematrix Interactive Entertainment System
  http://www.cinematrix.com/whatis.html

Fun and Games

• The BubbleGum Sequencer: a physical step sequencer that lets you create drumloops by arranging gumballs on a surface
  http://www.backin.de/gumball/
• Sony EyeToy: a vision-based "controller" allowing motion and gesture control of a variety of games
  http://www.eyetoy.com/
• Kinetic Combat: An eyetoy Martial Arts training and fitness game
  http://www.eyetoykinetic.com/
• Lego Mindstorms Vision Command: various projects using Lego and computer vision, including a “Party Optimizer” and a robotic xylophone
  http://mindstorms.lego.com/eng/products/vc/vcuserpreviews.asp
• The KidsRoom is an "Interactive Narrative Playspace" that uses three cameras for position and motion tracking
  http://vismod.media.mit.edu/vismod/demos/kidsroom/kidsroom.html
• Kickass Kung Fu is a video-game in which you use real martial-arts moves to control an on-screen kung-fu fighter in order to best both human and AI opponents
  http://www.kickasskungfu.net/en/
• The I/O Brush uses a camera to extract color, texture, and animation from the physical world and repaint it on a virtual canvas
  http://web.media.mit.edu/~kimiko/iobrush/
• Virtual Graffiti Wall using IR vision tracking
  http://www.elasticmind.com/LSC1/VGDemo.wmv
• Intelligent vision-based automatic pet door
  http://www.quantumpicture.com/Flo_Control/flo_control.htm
• Instant Shinerification: Pigeon Webcam dynamically adds virtual fez
  http://www.boingboing.net/2007/05/12/pigeon_webcam_dynami.html

This section contains links to tutorials and reference materials for many of the tools used in CS377S.

Tools for Rapid Prototyping

• Processing is a programming language and environment for programming images, animation, and sound. It is tailored towards artists and designers, and includes a capture library for grabbing video data from a camera.
  http://processing.org/
  http://processing.org/reference/libraries/video/index.html
• Myron (WebCamXtra) is a computer vision tool for artists that can be used in conjunction with Processing or with Macromedia Director. It includes color, motion, and “glob” tracking.
  http://webcamxtra.sourceforge.net/
• The Papier-Mâché toolkit, built on top of the Java Media Framework, includes vision-based object recognition.
  http://hci.stanford.edu/research/papier-mache/
• Eyepatch is a tool for visual, example-based training of computer vision classifiers.

Computer Vision Hardware

• Manu Kumar wrote a handy guide to modifying a Logitech webcam to work in the IR spectrum.
• The Nintendo Wiimote has a built-in IR camera, and you can read the data from the camera over Bluetooth. Check out Johnny Lee's Wiimote Tracking project page for videos and software.

Advanced Computer Vision Tools

• Matlab is data-manipulation software package that allows data to be analyzed and visualized using built-in functions and user-designed programs. Matlab's Image Processing Toolbox contains many useful functions for computer vision. Various Matlab resources are available online:
  • Matlab basics
    http://www.engin.umich.edu/group/ctm/basic/basic.html
  • Matlab Primer (PDF) from University of Florida
    http://web.mit.edu/6.777/www/downloads/primer.pdf
  • Matlab Tutorial (HTML) from University of Florida
    http://www.math.ufl.edu/help/matlab-tutorial/index.html
  • Matlab Tutorial from Utah College of Mines
    http://www.mines.utah.edu/gg_computer_seminar/matlab/matlab.html
  • MathWorks MATLAB Tutorial
    http://www.mathworks.com/academia/student_center/tutorials/launchpad.html
  • Online Matlab Help
    http://www.mathworks.com/access/helpdesk/help/techdoc/matlab.shtml
  • Matlab Central contains a user-contributed code library with many useful Downloadable Matlab code snippets.
    http://www.mathworks.com/matlabcentral/fileexchange/
  • Steve on Image Processing is a blog about image processing in Matlab.
    http://blogs.mathworks.com/steve/
• DirectShow is a Microsoft software component for capturing and displaying digital video. Though difficult to learn, it is the most flexible and powerful video capture solution for Windows platforms.
  • Flipcode has a throrough introductory tutorial in three parts.
    http://www.flipcode.com/articles/article_directshow01.shtml
  • Ruigang Yang has tutorial slides and several code examples.
    http://vis.uky.edu/~ryang/Miscellaneous/DShow/
• The Java Media Framework (JMF) is an optional Java package that extends the Java platform with audio and video capabilities. The JMF can capture, playback, stream, and transcode multiple media formats.
  • JMF programmer's guide
    http://java.sun.com/products/java-media/jmf/1.0/guide/index.html
  • JMF video capture example
    http://java.sun.com/products/java-media/jmf/2.1.1/solutions/JVidCap.html
• OpenCV is a powerful open-source C/C++ computer vision library. Many OpenCV tutorials are available online:
  • David Stavens runs an annual introduction to OpenCV for CS223B.
    http://ai.stanford.edu/~dstavens/cs223b/
  • The OpenCV Sourceforge repository includes the materials from a course on OpenCV at CVPR 2001.
    http://sourceforge.net/project/showfiles.phpgroup_id=22870&package_id=36092
  • Robert Laganière has a well-written but somewhat out-of-date guide to using OpenCV with MFC.
    http://www.site.uottawa.ca/~laganier/tutorial/opencv+directshow/cvision.htm
  • The OpenCV Yahoo Group is an excellent resource for finding answers to questions about OpenCV.
    http://tech.groups.yahoo.com/group/OpenCV/