OpenGM Benchmark (CVPR 2013)

Benchmark database of discrete energy minimization problems. For further details see

Models

dataset (zip)     model description (pdf)     evaluation (html)    

---

			Variables	Labels	Order	Structure	Functions	Instances	Reference	Comment
	In-Painting (N4) J. Lellmann et.al. converted by J. Lellmann and J.H. Kappes		14400	4	2	grid4	potts	2	[56]
	In-Painting (N8) J. Lellmann et.al. converted by J. Lellmann and J.H. Kappes		14400	4	2	grid8	potts	2	[56]
	ColorSegmentation (N4) J. Lellmann et.al. converted by J. Lellmann and J.H. Kappes		76800	3-12	2	grid4	potts	2	[56]
	ColorSegmentation (N8) J. Lellmann et.al. converted by J. Lellmann and J.H. Kappes		76800	3-12	2	grid8	potts	2	[56]
	ColorSegmentation K. Alahari et.al. converted by J.H. Kappes		21000-424720	3-4	2	grid8	potts	2	[19]
	Object Segmentation K. Alahari et.al. converted by J.H. Kappes		68160	4-8	2	grid4	potts	2	[19]
	MRF Photomontage R. Szeliski et.al. converted by J.H. Kappes		425632-514080	5-7	2	grid4	explicite	2	[64]	Models include soft-constraints.
	MRF Stereo R. Szeliski et.al. converted by J.H. Kappes		~100000	16-60	2	grid4	trunc. conv.	3	[64]
	MRF Inpainting R. Szeliski et.al. converted by J.H. Kappes		21838-65536	256	2	grid4	trunc. L2	2	[64]
	Chinese Characters S. Nowozin et.al. converted by S. Nowozin and J. H. Kappes		4992-17856	2	2	sparse	explicit	100	[58]
	Brain J. H. Kappes et.al. converted by J. H. Kappes		785540-7109137	5	2	3d-grid6	potts	5	[1]	Good example for large scale
	Scene Decomposition Gould et.al. converted by S. Nowozin and J. H. Kappes		150-208	8	2	sparse	explicit	751	[37]	Use superpixels.
	Geometric Surface Labeling (3) Gallagher et.al. converted by D. Batra and J. H. Kappes		29-1133	3	2	sparse	explicit	300	[35,38]	Use superpixels.
	Geometric Surface Labeling (7) Gallagher et.al. converted by D. Batra and J. H. Kappes		29-1133	7	2	sparse	explicit	300	[35,38]	Use superpixels.
	Matching N. Komodakis et.al. converted by N. Komodakis and J.H. Kappes		19-21	19-21	2	full / sparse	explicit	4	[54]	Local polytope relaxation is not tight.
	Cell Tracking B. X. Kausler et.al. converted by B. X.Kausler		41134	2	9	sparse	explicit	1	[46]	Model includes soft-constraints.
	Image Segmentation B. Andres et.al. converted by B. Andres		156-3764	156-3764	2	sparse	potts	100	[20]	Invariant to label permutation. Use superpixels.
	Hirachical Image Segmentation K. Sungwoong et.al. converted by K. Sungwoong and J. H. Kappes		122-651	122-651	34-651	sparse	potts	715	[48]	Invariant to label permutation. Use superpixels.
	3D Neuron Segmentation B. Andres et.al. converted by B. Andres		7958-101220	7958-101220	2	sparse	potts	2	[23,22]	Invariant to label permutation. Use supervoxels.