The Prague Texture Segmentation Datagenerator and Benchmark

The Prague Texture Segmentation Datagenerator and Benchmark - Result Details

optimized
for

It is possible that the resulting index maps don't correspond to the training set's indeces according to the data.xml file!
It leads to the wrong values of pixel-wise criteria as a result.


Benchmark - Colour [normal] CL		result by		veselj38's Histogram ratio features (1.2)

Note: TrainImageDivisionRatio=16, SegmentedPartSize=16


set \ criterion	CS	OS	US	ME	NE	O	C	CA	CO	CC	I	II	EA	MS	RM	CI	GCE	LCE	BCE	GBCE	BGM	SC	SSC	VD	L	AVI	NVI	NMI	M	ARI	JC	DC	FMI	WI	WII	NBDE
#3 - all	96.20	0.00	0.00	0.00	0.00	99.99	99.99	0.09	0.16	0.20	99.84	35.15	0.18	-31.69	3.69	0.18	7.26	6.45	8.10	7.29	96.20	92.71	92.70	3.80	96.10	2.65	15.05	84.88	4.79	89.27	86.71	92.88	92.88	92.75	93.01	4.31
#4 - all	18.66	0.00	0.00	77.59	76.87	98.94	98.89	23.60	24.15	34.72	75.85	17.34	28.47	-4.02	7.69	28.95	31.85	24.37	39.35	31.87	79.71	66.65	66.41	20.29	78.94	9.26	41.69	57.55	17.33	55.28	50.41	67.03	67.04	67.67	66.40	11.48
#5 - all	60.39	0.00	34.48	0.00	1.00	100.00	100.00	17.37	24.78	17.53	75.22	10.40	20.54	-8.59	14.49	20.85	8.92	4.99	25.31	21.37	85.18	81.10	77.89	9.63	79.62	4.56	17.66	79.74	8.42	79.06	73.66	84.83	85.13	78.28	92.58	4.29
#6 - all	0.00	0.00	0.00	100.00	100.00	99.99	99.93	0.34	2.15	0.49	97.85	16.32	0.66	-42.12	27.86	0.93	26.78	16.63	61.68	51.53	54.99	43.39	41.69	31.23	61.98	11.67	40.64	45.03	39.59	21.05	30.83	47.13	50.17	35.09	71.74	11.87
#7 - all	0.00	0.00	0.00	100.00	100.00	99.29	99.20	0.86	2.00	1.81	98.00	8.44	1.68	-36.63	13.38	1.78	44.12	36.85	60.93	53.66	57.38	43.45	44.29	36.76	53.52	14.65	46.97	45.39	22.24	33.16	30.47	46.71	47.39	39.94	56.23	9.27
#8 - all	32.21	0.00	0.00	63.87	59.49	100.00	100.00	12.04	18.91	15.61	81.09	10.69	16.52	-19.43	9.07	16.87	38.57	33.08	49.21	43.71	67.26	55.08	53.19	28.20	61.58	12.18	36.53	60.84	12.37	53.43	43.59	60.72	60.91	56.24	65.97	10.91
#9 - all	30.24	44.22	0.00	42.32	38.70	100.00	100.00	3.44	5.81	7.25	94.19	10.68	5.87	-27.99	6.93	6.19	35.03	25.36	51.87	42.20	63.44	51.48	54.49	26.95	53.67	11.86	33.68	62.17	13.07	52.82	43.48	60.61	60.76	65.27	56.57	12.60
#10 - all	0.00	17.89	0.00	81.45	78.31	100.00	100.00	1.46	2.47	3.08	97.53	11.01	2.64	-40.62	4.89	2.71	37.91	29.60	55.77	47.47	60.52	50.70	47.68	30.28	58.56	12.32	33.38	64.11	11.53	48.44	37.89	54.95	55.30	49.39	61.92	9.51
#11 - all	0.00	7.05	0.00	92.82	92.25	99.69	98.06	16.79	21.25	25.74	78.75	5.42	21.67	-15.70	7.92	22.51	37.32	26.32	51.77	40.76	67.52	56.61	53.95	26.40	64.10	11.74	30.54	67.54	9.73	53.16	41.51	58.66	58.69	57.03	60.39	7.63
#12 - all	0.00	0.00	0.00	100.00	100.00	100.00	100.00	16.53	19.49	22.99	80.51	6.99	21.05	-17.64	6.20	21.14	42.08	32.59	57.01	47.51	59.62	46.07	48.08	34.27	50.40	12.95	32.50	62.40	12.32	46.43	36.53	53.51	53.59	50.68	56.67	9.51
#6 - bark	23.51	0.00	32.97	35.26	29.96	99.36	89.80	3.89	11.33	10.59	88.67	8.16	6.81	-20.26	18.15	8.32	30.50	26.94	53.00	49.44	61.28	54.66	47.54	28.60	49.12	11.69	40.71	49.79	21.95	44.58	41.75	58.90	59.88	49.94	71.79	10.10
#6 - flowers	47.27	24.47	0.00	44.36	43.70	97.14	98.86	7.52	13.75	10.28	86.25	16.54	11.67	-27.23	9.20	11.84	37.80	30.68	46.13	39.02	65.45	55.30	57.26	28.94	60.88	11.70	40.73	56.07	14.13	57.32	49.54	66.26	66.26	65.34	67.19	5.98
#6 - glass	17.49	0.00	20.69	58.89	57.86	100.00	100.00	0.20	0.36	0.46	99.64	20.00	0.40	-38.82	7.57	0.40	30.38	22.15	44.46	36.23	73.12	61.92	61.00	23.20	70.41	9.38	32.64	64.22	14.59	55.34	47.62	64.52	64.52	64.56	64.48	10.40
#6 - man-made	59.08	0.00	0.00	35.64	33.21	100.00	100.00	21.56	24.76	22.82	75.24	12.31	23.73	0.73	6.03	23.76	24.09	20.33	32.45	28.68	80.75	73.52	69.99	18.03	73.86	8.32	28.97	68.63	9.67	70.94	62.68	77.06	77.10	74.59	79.69	7.89
#6 - nature	44.87	50.05	0.00	41.65	46.85	88.22	97.49	14.37	17.76	20.57	82.24	24.91	18.23	-18.26	24.13	18.64	27.50	20.36	47.90	40.76	67.85	53.14	61.38	23.65	53.36	10.16	35.38	54.66	20.17	55.42	53.93	70.07	70.96	83.17	60.54	14.07
#6 - plants	0.00	0.00	24.54	74.65	69.34	100.00	100.00	5.02	7.94	7.81	92.06	15.16	7.87	-22.76	7.19	7.87	44.00	38.15	64.50	58.66	43.34	35.92	34.68	43.26	52.42	14.34	49.93	45.22	23.85	28.32	27.52	43.17	43.45	38.77	48.68	17.24
#6 - rock	0.00	0.00	0.00	100.00	100.00	58.60	67.37	28.98	43.82	42.85	56.18	9.50	39.51	19.25	11.95	41.28	43.67	36.59	57.31	50.23	57.63	43.64	42.96	36.81	58.30	12.96	45.13	51.70	20.43	36.99	33.14	49.78	49.83	47.81	51.93	11.39
#6 - stone	0.00	45.85	0.00	53.07	51.82	87.76	74.23	33.86	37.48	76.11	62.52	3.62	45.98	15.11	17.68	50.84	24.40	15.41	62.11	53.12	50.75	40.28	45.77	30.35	50.08	11.39	39.67	54.24	24.30	39.02	36.20	53.15	56.23	78.90	40.08	10.13
#6 - textile	13.20	0.00	0.00	85.42	86.37	95.07	97.23	15.04	21.64	27.84	78.36	11.98	21.72	-8.26	11.90	23.09	44.85	34.10	57.57	46.82	55.31	43.34	43.65	36.46	55.66	12.85	44.72	51.79	21.58	34.44	31.63	48.06	48.07	48.85	47.30	13.88
#6 - wood	0.00	0.00	0.00	100.00	100.00	100.00	100.00	17.55	20.51	25.96	79.49	10.31	22.38	-7.86	7.89	22.79	36.55	26.78	47.29	37.52	71.86	56.59	59.65	27.21	60.53	10.40	36.20	57.99	17.73	50.71	45.24	62.30	62.31	63.83	60.84	12.99
mean	22.16	9.48	5.63	64.35	63.29	96.20	96.05	12.03	16.03	18.74	83.97	13.25	15.88	-17.64	11.19	16.55	32.68	25.38	48.69	41.39	65.96	55.28	55.21	27.22	62.15	10.85	36.14	59.20	16.99	50.26	45.22	61.01	61.52	60.41	63.70	10.27


criterion	Graph
correct
over
under
missed
noise
F-measure


image	Texture mosaic (undegraded)	Ground truth	Segmentation
#3 - all
#4 - all
#5 - all
#6 - all
#7 - all
#8 - all
#9 - all
#10 - all
#11 - all
#12 - all
#6 - bark
#6 - flowers
#6 - glass
#6 - man-made
#6 - nature
#6 - plants
#6 - rock
#6 - stone
#6 - textile
#6 - wood