Master of Computer Science
Department of Computer Science
New York University
Email
wch225@cs.nyu.edu
Advanced Topics in Computer Vision and Tracking - Spring 2006
Advanced
Topics in Computer Vision and Tracking Prof. Bregler
Homework assignment:
Assignment
0
Assignment
1 Implement and experiment with "Lucas-Kanade"
ShiTomasi94
Bouget00
Assignment
2 Implement and experiment with a Skin-Color Tracker
Papers on nonlinear Spaces (for Mar-07 class):
LLE
SVM
Paper Presentation:
Real
Time Face and Object Tracking as a Component of a Perceptual
User Interface
Gary R. Bradski, Intel Corporation, Microcomputer Research Lab.
gary.bradski@intel.com
Final Project:
Recognising
Panoramas
M. Brown and D. G. Lowe Department of Computer Science, University of
British Columbia, Vancouver, Canada.
The problem considered in this paper is the fully automatic construction of
panoramas. Fundamentally, this problem requires recognition, as we need to
know which parts of the panorama join up. Basically, there are two
parts of method for automatic image matching: direct and feature
based. Feature based methods begin by establishing correspondences between
points, lines or other geometrical entities. In this project, I will
try to use matlab to implement two kinds of feature based algorithm, Scale Invariant Feature
Transform (SIFT) and extracting Harris corners and using
a normalised cross-correlation of the
local intensity values.
Result
Original Image
: This is the image using in the Lucas-Kanade
assignment
top of the original
image : This is the image uesd to demo the algorithm
down of the original
image : This is the image uesd to demo the algorithm
image after feature matching by extracting Harris corners and using
a normalised cross-correlation of the local intensity values (click for big
image)
image after feature matching by extracting Harris corners and using
a normalised cross-correlation of the
local intensity values. (click for big
image)
image after feature matching by Scale Invariant Feature
Transform (SIFT) (click for big image)
image after feature matching by Scale Invariant Feature
Transform (SIFT) (click for big image)
image after feature matching (click for big image)
Conclusion
Although I've done the Feature Matching and Image Matching, there are still two parts, Bundle Adjustment and
Multi-band Blending, and other works need to done.
Really thanks for Prof Chris Bregler
