Graph Databases
Graph formats
The general/standard format to store graphs is GXL (Graph eXchange Language), which is an extensible markup language (XML) .
However, in this work all provided source codes and binaries rely on a lighter representation format for graphs, referred to as DAT. The reason behind using this format is that all methods are developed in C, and with such format it is easier/faster to parse graphs.
More details about DAT format, and a code snippet can be found in the Source Code page.
Every database mentioned below is given in the two versions (GXL and DAT).
Please refer to the manuscript (Section 4.2) for details and explanations about each database and its cost functions.
For certain databases, subsets and cost functions are provided as well.
Databases
MUTA
Download here.
Ref. [Abu-Aisheh et al. (2015). A graph database repository and performance evaluation metrics for graph edit distance. IAPR]
PROTEIN
Download here.
Ref. [Riesen et al. (2008). IAM graph database repository for graph based pattern recognition and machine learning. S+SSPR]
PAH
Download here.
Ref. [Abu-Aisheh et al. (2015). A graph database repository and performance evaluation metrics for graph edit distance. IAPR]
CMU-HOUSE
There are two versions of this DB:
- Default - vertices are in their default order
- Randomized - vertices are re-ordered randomly
Download here.
Ref. [Moreno-Garcia et al. (2016). A graph repository for learning error tolerant graph matching. S+SSPR]
CMU-HOTEL
Download here.
Ref. [Moreno-Garcia et al. (2016). A graph repository for learning error tolerant graph matching. S+SSPR]
VOC-CAR
Download here.
Ref. [Zhou et al. (2012). Factorized graph matching. CVPR]
MNIST
Download here.
Ref. [Cecotti. (2016).Active graph based semi-supervised learning using image matching: application to handwritten digit recognition. PRL]
LETTER
Download here.
Ref. [Riesen et al. (2008). IAM graph database repository for graph based pattern recognition and machine learning. S+SSPR]
PALMPRINT
Download here.
Ref. [Han et al. (2007). Palmprint recognition based on directional features and graph matching. ICB]
SYNTHETIC
Download here.
(The script to generate the graphs is provided as well in the compressed file)Ref. [Csardi et al. (2006). The igraph software package for complex network research. IJCS]