A Comparative Study of Graph Matching Algorithms in Computer Vision

Benchmark Results for “car”

This page shows the benchmarks results for the dataset “car”. The reported values, bounds and accuracies are averaged across all instances of the dataset.

Run time 1s

method avg value avg bound feasible optimal accuracy
dd-ls0 -69.3346 -69.4221 30 / 30 29 / 30 91.15%
dd-ls3 -56.8328 -74.7414 30 / 30 14 / 30 74.02%
dd-ls4 -49.3798 -78.2581 30 / 30 1 / 30 58.51%
fgmd inf -inf 0 / 30 0 / 30
fm-bca -69.3578 -70.6514 30 / 30 28 / 30 91.66%
fm -68.9494 -106.915 30 / 30 23 / 30 87.97%
fw -63.0926 -inf 30 / 30 2 / 30 62.61%
ga -68.4517 -inf 30 / 30 17 / 30 83.58%
hbp inf -inf 23 / 30 23 / 30
ipfps -65.2093 -inf 30 / 30 3 / 30 79.67%
ipfpu -60.2257 -inf 30 / 30 2 / 30 68.81%
lsm -51.2689 -inf 30 / 30 0 / 30 51.89%
mp -69.048 -70.5785 30 / 30 24 / 30 92.04%
mp-fw -69.2954 -70.6551 30 / 30 27 / 30 90.64%
mpm inf -inf 23 / 30 2 / 30
mp-mcf -69.0429 -70.4127 30 / 30 26 / 30 90.75%
pm -34.6061 -inf 30 / 30 0 / 30 23.03%
rrwm -68.0112 -inf 30 / 30 11 / 30 86.86%
sm -62.7421 -inf 30 / 30 2 / 30 75.85%
smac -52.1913 -inf 30 / 30 0 / 30 52.11%

Run time 10s

method avg value avg bound feasible optimal accuracy
dd-ls0 -69.3352 -69.4186 30 / 30 29 / 30 91.15%
dd-ls3 -67.6957 -69.7691 30 / 30 26 / 30 89.59%
dd-ls4 -59.4165 -73.0657 30 / 30 17 / 30 77.28%
fgmd inf -inf 29 / 30 25 / 30
fm-bca -69.3611 -70.646 30 / 30 29 / 30 91.76%
fm -69.0768 -106.915 30 / 30 27 / 30 89.02%
fw -63.0926 -inf 30 / 30 2 / 30 62.61%
ga -68.4517 -inf 30 / 30 17 / 30 83.58%
hbp inf -inf 28 / 30 26 / 30
ipfps -65.2093 -inf 30 / 30 3 / 30 79.67%
ipfpu -60.2257 -inf 30 / 30 2 / 30 68.81%
lsm -51.2689 -inf 30 / 30 0 / 30 51.89%
mp -69.1376 -70.5569 30 / 30 25 / 30 92.42%
mp-fw -69.3619 -70.1549 30 / 30 29 / 30 91.43%
mpm -56.9838 -inf 30 / 30 2 / 30 64.89%
mp-mcf -69.1532 -69.9693 30 / 30 27 / 30 91.13%
pm -34.6061 -inf 30 / 30 0 / 30 23.03%
rrwm -68.0112 -inf 30 / 30 11 / 30 86.86%
sm -62.7421 -inf 30 / 30 2 / 30 75.85%
smac -52.1913 -inf 30 / 30 0 / 30 52.11%

Run time 100s

method avg value avg bound feasible optimal accuracy
dd-ls0 -69.3352 -69.4186 30 / 30 29 / 30 91.15%
dd-ls3 -69.363 -69.4026 30 / 30 29 / 30 91.15%
dd-ls4 -67.4528 -69.8103 30 / 30 26 / 30 89.40%
fgmd -69.0821 -inf 30 / 30 25 / 30 89.09%
fm-bca -69.3613 -70.646 30 / 30 29 / 30 91.66%
fm -69.0768 -106.915 30 / 30 27 / 30 89.02%
fw -63.0926 -inf 30 / 30 2 / 30 62.61%
ga -68.4517 -inf 30 / 30 17 / 30 83.58%
hbp -69.2069 -72.7786 30 / 30 26 / 30 90.51%
ipfps -65.2093 -inf 30 / 30 3 / 30 79.67%
ipfpu -60.2257 -inf 30 / 30 2 / 30 68.81%
lsm -51.2689 -inf 30 / 30 0 / 30 51.89%
mp -69.2371 -70.5566 30 / 30 26 / 30 92.13%
mp-fw -69.3712 -69.6033 30 / 30 30 / 30 91.35%
mpm -56.9838 -inf 30 / 30 2 / 30 64.89%
mp-mcf -69.2868 -69.5731 30 / 30 28 / 30 91.23%
pm -34.6061 -inf 30 / 30 0 / 30 23.03%
rrwm -68.0112 -inf 30 / 30 11 / 30 86.86%
sm -62.7421 -inf 30 / 30 2 / 30 75.85%
smac -52.1913 -inf 30 / 30 0 / 30 52.11%

Run time 300s

method avg value avg bound feasible optimal accuracy
dd-ls0 -69.3352 -69.4186 30 / 30 29 / 30 91.15%
dd-ls3 -69.363 -69.4026 30 / 30 29 / 30 91.15%
dd-ls4 -69.3389 -69.4115 30 / 30 29 / 30 91.83%
fgmd -69.0821 -inf 30 / 30 25 / 30 89.09%
fm-bca -69.3613 -70.646 30 / 30 29 / 30 91.66%
fm -69.0768 -106.915 30 / 30 27 / 30 89.02%
fw -63.0926 -inf 30 / 30 2 / 30 62.61%
ga -68.4517 -inf 30 / 30 17 / 30 83.58%
hbp -69.2069 -72.7786 30 / 30 26 / 30 90.51%
ipfps -65.2093 -inf 30 / 30 3 / 30 79.67%
ipfpu -60.2257 -inf 30 / 30 2 / 30 68.81%
lsm -51.2689 -inf 30 / 30 0 / 30 51.89%
mp -69.2371 -70.5565 30 / 30 26 / 30 92.13%
mp-fw -69.3712 -69.5464 30 / 30 30 / 30 91.35%
mpm -56.9838 -inf 30 / 30 2 / 30 64.89%
mp-mcf -69.3001 -69.5379 30 / 30 28 / 30 91.38%
pm -34.6061 -inf 30 / 30 0 / 30 23.03%
rrwm -68.0112 -inf 30 / 30 11 / 30 86.86%
sm -62.7421 -inf 30 / 30 2 / 30 75.85%
smac -52.1913 -inf 30 / 30 0 / 30 52.11%

Per Instance Results

Results for individual instances of the dataset are also available: