CREMI

Results for the neuron segmentation category, averaged over all samples.

Group	Submission	CREMI score	VOI split	VOI merge	ARAND
SeungLab-eding	Seung-ed	1.525	0.895	2.160	0.764
afish1001	OurUnet	2.197	0.698	5.034	0.855
afish1001	ABWUnet	2.199	0.696	5.042	0.856
HVCL	Deep Res	2.212	1.672	4.182	0.837
afish1001	wunet_wj	2.212	0.641	5.157	0.854
afish1001	unet_wj	2.223	0.648	5.178	0.858
afish1001	wunet_ce	2.244	0.829	4.982	0.882
afish1001	unet_ce	2.246	0.831	4.987	0.883
VCG	testmala	2.412	3.297	3.283	0.888
IIL+DIVE@TAMU	new_segm	2.671	0.928	6.475	0.964
VCG	ZWZ-0.0	2.675	3.524	4.193	0.927
-	testsubm	2.683	3.083	4.768	0.918
Cvlab	cvlab_2	2.756	1.526	6.411	0.957
Cvlab	cvlab_1	2.891	4.686	5.421	0.828
-	Final	2.896	3.805	5.599	0.895
DIVE	BasUnet	2.959	1.740	7.366	0.962
DIVE	RUseg	2.976	6.003	3.121	0.972
Cvlab	cvlab	3.472	6.069	6.542	0.956

CREMI score: The geometric mean of (VOI split + VOI merge) and ARAND.
VOI split and merge: The Variation of Information between a segmentation X and ground truth Y. The split and merge parts correspond to the conditional entropies H(X|Y) and H(Y|X), respectively.
ARAND: The Adapted Rand Error, i.e., 1.0 - Rand F-Score.

For each value, lower is better.

Results for the synaptic cleft detection category, averaged over all samples.

Group	Submission	CREMI score	FP	FN	1 - F-score	ADGT	ADF
DIVE	Rclefts	380.86	167900713.3	0.0	0.990	761.72	0.00
DIVE	OrClefts	380.86	167900713.3	0.0	0.990	761.72	0.00
DIVE	ResUnet	380.86	167900713.3	0.0	0.990	761.72	0.00
CMM	GenT	380.86	167900713.3	0.0	0.990	761.72	0.00
DIVE	CRunet2	318.85	107459361.0	4195.7	0.824	611.30	26.40
DIVE	ex19t5	315.97	64192.0	1757.3	0.112	600.79	31.15
DIVE	exp18_la	284.83	684532.0	0.0	0.493	564.72	4.94
DIVE	ex19t2	283.68	173078.7	0.0	0.201	551.77	15.60
cxj	USTC	284.98	1015564.7	108523.3	0.600	516.31	53.65
DIVE	Exp12_64	202.46	335698.7	0.0	0.374	402.99	1.93
DIVE	cleft_A+	199.35	378215.7	0.0	0.386	395.99	2.72
DIVE	3class8t	198.78	262876.7	0.0	0.282	393.81	3.76
3DEM	ABC_C	375.91	289364.7	89797.7	0.619	385.49	366.34
DIVE	cleftlol	179.28	723352.7	2716.7	0.375	334.05	24.51
DIVE	CRunet22	180.77	162763.0	23055.7	0.312	305.59	55.95
DIVE	SegClef1	156.70	617804.0	3339.3	0.373	296.85	16.56
DIVE	cleft_lb	139.19	284326.0	0.0	0.349	265.87	12.52
IAL	3d_model	229.78	86415.3	84295.0	0.317	255.27	204.29
DIVE	exp18_be	136.63	8797.3	0.0	0.014	247.19	26.08
VCG	FgRj0805	117.25	1329995.3	327.7	0.610	232.66	1.84
VCG	FgRj0808	117.25	1329995.3	327.7	0.610	232.66	1.84
IAL	A+test	153.91	196558.0	55962.3	0.339	232.28	75.53
IAL	unaries	150.64	175624.3	60151.3	0.327	213.52	87.77
SDG	SnapX.O	124.59	182577.0	17000.7	0.306	205.18	44.01
SDG	SnapF.V	124.59	182577.0	17000.7	0.306	205.18	44.01

CREMI score: The mean of ADGT and ADF. We are not including the F-Score for now, as our current way of computing it can lead to unfair comparison. We are working on a more robust measure and will update the results accordingly.
FP, FN, and F-Score: The False Positives, False Negatives, and the resulting F-Score in the clefts detection volume. See metrics for details.
ADGT: The average distance of any found cleft voxel to the closest ground truth cleft voxel.
ADF: The average distance of any ground truth cleft voxel to the closest found cleft voxel.

For each value, lower is better (F-Score shown as 1 - F-Score).

Results for the synaptic partner identification category, averaged over all samples.

Group	Submission	CREMI score	FP	FN
PNI	asyn_mod	0.423	367.667	262.333
IAL	PSP_unar	0.461	266.667	281.000
HCBS	Tr66_80K	0.449	223.000	286.667
HCBS	Tr66comb	0.451	219.000	292.333
PNI	asyn_at0	0.468	293.333	297.000
PNI	asyn_ori	0.493	310.000	302.333
IAL	PSP_full	0.464	187.333	310.000
NO2	lr_balan	0.447	175.333	314.000
PNI	asyn_at1	0.490	274.333	314.667
HCBS	Tr66t-01	0.469	182.667	318.667
HCBS	PrnTrn66	0.465	178.667	322.333
DIVE	PTR_V2	0.453	162.667	333.667
NO2	lr_affin	0.567	118.333	450.000
AnonumoysGroup	PrTrn66	0.620	238.000	451.667

CREMI score: 1 - F-Score of the FP and FN.
FP, FN: The False Positives and False Negatives synaptic partner pairs. See metrics for details.

For each value, lower is better.

MICCAI Challenge on Circuit Reconstruction from Electron Microscopy Images