Fast Affine Template Matching over Galois Field

Chao Zhang1  Takuya Akashi2  
Smart Computer Vision Laboratory (Akashi Labo)
Iwate University, Japan


Benchmark (500 images and ground truth, 24.9 MB)          Our results for easy comparison (9.0 KB)  

Method overview

 

 

Abstrat
In this paper, we address the problem of template matching under affine transformations with general images. Our approach is to search an approximate affine transformation over a binary Galois field. The benefit is that we can avoid matching with huge amount of potential transformations, because they are discretely sampled. However, a Galois field of affine transformation can still be impractical for exhaustive searching. To approach the optimum solution efficiently, we introduce a level-wise adaptive sampling (LAS) method under genetic algorithm framework. In LAS, individuals converge to the global optimum according to a level-wise selection and crossover while the population number is decreased by a population bounding scheme. In the experiment section, we analyse our method systematically and compare it against the state-of-the-art method on an evaluation data set. The results show that our method has a high accuracy performance with few matching tests compared against the state-of-the-art method.

 

 

 

README
./benchmark/targets: This folder contains all the target/source images.
./benchmark/templates: This folder contains all the rectangular template images.
./file_ID.txt: This file records the filename of each target image for easy programming.
./GT_affine.yml: This file records the groud truth 3x3 affine matrix of each matching.
./GT_coordinate.yml: This file records the four points of each ground truth area.
./GT_homography.yml: This file records the homography matrix. When generating each template, findHomography function is used to find the warp matrix.
./GT_SAD.yml: This file records the SAD error of each ground truth caused by interpolation.
./compare.txt: This file records the matching times and overlap error of both our method and the comparative method (Fast-Match).
./compare.py: This is a python script for generating the figure in our paper. (overlap error)
./compare1.py: This is a python script for generating the figure in our paper. (matching times)

 

 

 

Citation
If you find this dataset useful, please cite the following paper:
@conference{zhang2015matching,
title={Fast Affine Template Matching over Galois Field},
author={Zhang, Chao and Akashi, Takuya},
booktitle={British Machine Vision Conference (BMVC)},
year={2015}
}

 

 

 

Copyright

THIS DATA SET IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.

The images provided above may have certain copyright issues. We take no guarantees or responsibilities, whatsoever, arising out of any copyright issue. Please use it at your own risk.