>
ANG LEE
SOFTWARE ENGINEER, MAGICIAN
http://anglee.org
ang.lee[at]gmail.com

Projects::Feature Matching for Location/Object Recognition

Given a collection of images(library), this project is about finding the one that contains the same feature as a query image that is not in the library. SIFT was used to identify features and then similarity between features is computed.

Task

The task is to find the closest image in the library that matches a particular query image.

Input

A collection of images (the library), 31 photos from a trip to Venice, Italy. Click on a image to enlarge.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

 

Query Images, Click on a image to enlarge.

 

Results:

Query Image #1 Best Match, Image#3, vote = 249
2nd place, Image#30, vote = 186
Histogram (click to enlarge) 3rd place, Image#24, vote = 179

 

Query Image #2 Best Match, Image#9, vote = 120
2nd place, Image#25, vote = 119
Histogram (click to enlarge) 3rd place, Image#26, vote = 118

 

Query Image #3 Best Match, Image#18, vote = 128
2nd place, Image#30, vote = 110
Histogram (click to enlarge) 3rd place, Image#26, vote = 102

 

Query Image #4 Best Match, Image#18, vote = 102
2nd place, Image#30, vote = 93
Histogram (click to enlarge) 3rd place, Image#13, vote = 87

 

Query Image #5 Best Match, Image#1, vote = 135
2nd place, Image#30, vote = 96
Histogram (click to enlarge) 3rd place, Image#24, vote = 96

 

 

Links:

Original project page

Assignment Instruction

CS7495 2005 Class Website

SIFT Info


Last update: 1/20/2009
Ang Lee 4371 Winters Chapel Rd., Apt#3226 Atlanta, GA 30360, USA
(C) Copyright, Ang Lee, 2009, All rights reserved.