Skip to content

The C++ project for traditional image keypoint detectors and descriptors

License

Notifications You must be signed in to change notification settings

wlzhao22/lip-vireo

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

58 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

lip-vireo

This is a C++ project that imcludes the implementations of the traditional image keypoint detectors and descriptors.

Overview

Nowadays, traditional image keypoint detectors and descriptors are no longer popular as they were in 2000s. They have given place to the deep features, which become increasingly popular since 2012. However, the theories upon which the traditional detectors and descriptors are built are still shiny in nowadays. The scale-space theory from Dr. Tony Lindberg is beautiful. I learned a lot from his papers and book. It took me hundreds of weekends to implement all the detectors and descriptors in this project. This project also saw the birth of 'Flip invariant SIFT', which is my proposal. It well addressed the sensitivity issue of SIFT to flip transformation. The project name was given by my PhD supervisor as this project was started during my PhD in VIREO research group in City University of Hong Kong.

The included detectors are Hessian, laplacian-of-Gaussian, Harris and SURF. The included descriptors are SIFT, LJET, Flip-invariant SIFT, RIFT, and SPIN. Please find out more details from the manual.

Detectors and Descriptors

Detectors: Difference-of-Gaussian (DoG), Laplacian-of-Gaussian (LoG), Hessian, Hessian-Laplacian, Harris, Harris-Laplacian, SURF, and Hessian-Affine

Descriptors: SIFT [1], F-SIFT [2], SURF [3], FIND, SPIN, LJET, and RIFT

[1] David G. Lowe: Distinctive Image Features from Scale-Invariant Keypoints, IJCV'04.
[2] Wan-Lei Zhao, Chong-Wah Ngo: Flip-invariant SIFT for Copy and Object Detection, TIP'13.
[3] Herbert Bay, Tinne Tuytelaars, Luc Van Gool: SURF: Speeded up robust features, ECCV'06

Compile

This project can be smoothly compiled under Ubuntu and MacOS, while it is also possible to be compiled successfully (by MinGW) under Windows with only minor changes. For Windows, user should install MinGW first, then edit "iotool.cpp" and "config.h", comment out the following lines in the files
#ifndef LINUX
#define LINUX
#endif

Step 1. Download libjpeg, zlib and libpng packages and compile them respectively, copy libjpeg.a, zlib.a and libpng.a to "libs/" folder

Step 2. Go to zlib folder and libpng folder respectively, run "make install" for zlib and libpng on your system

Step 3. Run "make release" under the folder as follows

$ cd lip-vireo
$ make release
The prerequisite packages could be found under the "packages" folder or from the following URLs

Manual

A PDF manual of the overview about the detectors and descriptors, as well as the call of lip-vireo from commandline could be found under "manual" folder.

Copyright Issues

1. DoG+SIFT has been patented by David G. Lowe in US (Patent No. US 6,711,293B1). The patent holder is University of British Columbia, Canada
2. Copyright holder for Fast Hessian and SURF descriptor are Tinne Tuytelaars and et al.;
3. Copyright holder for FIND descriptor are Mr. Xiao-Jie Guo and et al. from University of Tianjin, China.

Author

Wan-Lei Zhao
He has been benefit from many open-soure projects. He is willing to share his work as well.