The Prague Texture Segmentation Datagenerator and Benchmark - Results
optimized
for

Results are sorted by criterion 'CS' (correct segmentation) in descending order. The sorting criterium or the order can be changed by clicking on the appropriate criterium label. Subset of compared results can be filtered by several filters below. The segmentation details (single mosaics and their corresponding criteria values) are visible by clicking on the order number.

Select  result's set    benchmark     hidden results Show Hide
Filter by algorithm features
? 0 1   f1 = classification (supervised segmentation)
? 0 1   f2 = hiearchy result (manual selection)
? 0 1   f3 = known number of regions
? 0 1   f4 = [reserved]
Filter by degradationtype: ? no Gaussian Poisson Salt&pepper Blur (Gaussian) Median
Filter by algorithm
Algorithm: Version:

Benchmark - Colour f1 f2 f3 f4 RANK criterion 'CS' (correct segmentation) in descending orderCS OS US ME NE O C CA CO CC I. II. EA MS RM CI GCE LCE dD dM dVI
 1.  huangyuan's EWT-FCNT   [normal] CL 1 0 0 0  (1.29)  98.45
(1)
0.00
(1)
0.00
(1)
0.37
(2)
0.46
(2)
0.93
(1)
1.04
(1)
97.67
(1)
98.78
(1)
98.81
(1)
1.22
(1)
0.25
(2)
98.77
(1)
98.17
(1)
0.24
(2)
98.78
(1)
2.23
(1)
1.68
(1)
1.21
(1)
0.74
(1)
13.68
(3)
 2.  huangyuan's unet psp21ewt  [normal] CL 1 0 0 0  (2.48)  98.09
(2)
1.70
(11)
0.00
(1)
0.37
(1)
0.09
(1)
1.57
(2)
1.82
(2)
96.52
(2)
98.18
(2)
98.01
(2)
1.82
(2)
0.39
(3)
98.08
(2)
97.27
(2)
0.29
(3)
98.09
(2)
3.20
(2)
2.70
(2)
1.75
(2)
1.08
(2)
13.71
(4)
 3.  v-andrearczyk's FCNT sup.  [normal] CL 1 0 1 0  (3.52)  96.01
(3)
1.56
(10)
1.20
(3)
0.78
(3)
0.89
(3)
2.72
(3)
2.29
(3)
93.95
(3)
96.73
(3)
97.02
(3)
3.27
(3)
0.68
(4)
96.68
(3)
95.10
(3)
0.86
(4)
96.77
(3)
5.55
(3)
3.75
(3)
3.06
(3)
1.96
(3)
13.80
(5)
 4.  connetwork12's conn col-noblur  [normal] CL 1 0 0 0  (4.14)  84.57
(4)
0.00
(1)
1.70
(4)
9.50
(4)
10.22
(4)
7.00
(4)
5.34
(4)
86.21
(4)
92.02
(4)
92.68
(4)
7.98
(4)
1.70
(5)
91.72
(4)
88.03
(4)
2.08
(5)
92.02
(4)
11.76
(4)
8.61
(4)
7.50
(4)
4.69
(4)
13.99
(8)
 5.  connetwork12's conn col-blur3  [normal] CL 1 0 0 0  (5.76)  72.93
(5)
1.73
(12)
2.74
(8)
18.89
(5)
18.74
(5)
9.20
(5)
9.78
(5)
80.08
(5)
87.78
(5)
89.19
(5)
12.22
(5)
2.59
(6)
87.40
(5)
81.67
(5)
3.15
(6)
87.92
(5)
16.64
(5)
12.55
(5)
10.90
(5)
6.78
(5)
14.18
(9)
 6.  hadacmar's CoOccurenceFeatures  [normal] CL 1 0 1 0  (6.38)  52.48
(6)
0.00
(1)
1.94
(6)
41.55
(7)
40.97
(7)
20.74
(7)
22.10
(7)
67.01
(6)
77.86
(6)
78.34
(7)
22.14
(6)
4.40
(7)
76.21
(6)
66.79
(6)
4.47
(7)
77.05
(6)
23.94
(7)
19.69
(7)
17.86
(6)
10.62
(6)
14.22
(10)
 7.  munozfab's MRF Gibbs sampler  [normal] 1 0 1 0  (7.48)  46.11
(7)
0.81
(8)
4.18
(9)
44.82
(8)
45.29
(8)
14.52
(6)
16.77
(6)
65.42
(7)
76.19
(7)
80.30
(6)
23.81
(7)
4.82
(8)
75.40
(7)
64.29
(7)
6.43
(8)
76.69
(7)
25.79
(8)
20.68
(8)
20.35
(7)
13.25
(7)
14.51
(11)
 8.  yzan's Laws & Naive Bayes   [normal] CL 1 0 1 0  (11.71)  42.19
(8)
0.00
(1)
2.41
(7)
51.41
(9)
51.41
(9)
97.97
(15)
98.38
(15)
2.69
(15)
4.97
(15)
6.86
(15)
95.03
(15)
15.57
(15)
4.80
(15)
-34.39
(16)
12.79
(13)
5.29
(15)
30.12
(9)
23.82
(9)
23.85
(9)
14.07
(9)
14.73
(12)
 9.  borkojos's cooccurence&naive bayes   [normal] CL 1 0 1 0  (8.48)  39.12
(9)
0.00
(1)
1.92
(5)
54.14
(10)
54.79
(10)
26.75
(8)
29.46
(8)
59.27
(8)
70.76
(8)
75.10
(8)
29.24
(8)
4.85
(9)
70.29
(8)
56.14
(8)
6.65
(9)
71.51
(8)
32.19
(10)
25.91
(10)
24.01
(10)
14.22
(10)
14.75
(13)
f1 f2 f3 f4 RANK criterion 'CS' (correct segmentation) in descending orderCS OS US ME NE O C CA CO CC I. II. EA MS RM CI GCE LCE dD dM dVI
 10.  LiRa's ==   [normal] CL 1 0 0 0  (11.29)  29.72
(10)
41.22
(16)
6.48
(11)
27.12
(6)
29.19
(6)
100.00
(16)
100.00
(16)
0.00
(16)
0.00
(16)
0.00
(16)
100.00
(16)
0.00
(1)
0.00
(16)
0.00
(14)
0.00
(1)
0.00
(16)
19.97
(6)
13.74
(6)
23.19
(8)
13.59
(8)
17.36
(16)
 11.  leskajur's LFM + kNN   [normal] CL 1 0 1 0  (11.29)  26.00
(11)
5.01
(14)
9.42
(12)
60.96
(12)
59.65
(12)
66.89
(11)
60.33
(12)
31.40
(11)
43.31
(11)
48.73
(12)
56.69
(11)
11.45
(12)
39.22
(11)
14.96
(11)
13.51
(14)
41.76
(11)
33.28
(11)
27.61
(11)
30.99
(12)
23.69
(13)
13.67
(2)
 12.  luptarad's laws masks & kNN 1.0  [normal] CL 1 0 1 0  (10.76)  24.56
(12)
4.40
(13)
9.46
(13)
58.04
(11)
55.34
(11)
51.04
(9)
47.53
(10)
36.75
(9)
49.97
(10)
51.28
(10)
50.03
(10)
7.75
(10)
45.09
(10)
24.95
(10)
11.67
(11)
47.24
(10)
33.82
(12)
28.68
(13)
31.26
(13)
22.73
(12)
13.83
(7)
 13.  veselj38's Histogram ratio features 1.21  [normal] CL 1 0 0 0  (13.19)  20.64
(13)
8.84
(15)
5.46
(10)
62.70
(13)
62.78
(13)
95.42
(14)
94.85
(14)
9.45
(14)
13.85
(14)
17.65
(14)
86.15
(14)
13.72
(14)
13.61
(14)
-19.75
(15)
10.26
(10)
14.53
(14)
37.37
(13)
28.64
(12)
30.55
(11)
17.96
(11)
15.09
(15)
 14.  jirkujak's LBP 3,1  [normal] CL 1 0 1 0  (11.81)  12.16
(14)
0.00
(1)
9.67
(15)
75.55
(14)
74.24
(14)
72.24
(12)
51.36
(11)
26.31
(12)
40.38
(12)
51.72
(9)
59.62
(12)
10.11
(11)
35.36
(12)
10.56
(12)
17.37
(16)
39.04
(12)
37.52
(14)
31.64
(14)
37.57
(15)
30.94
(15)
13.56
(1)
 15.  guloleg's HOG  [normal] 1 0 1 0  (11.52)  11.59
(15)
1.54
(9)
9.62
(14)
76.51
(15)
77.17
(15)
51.11
(10)
46.26
(9)
35.67
(10)
52.72
(9)
49.37
(11)
47.28
(9)
12.13
(13)
47.43
(9)
29.09
(9)
12.53
(12)
49.14
(9)
37.70
(15)
32.93
(15)
36.90
(14)
27.03
(14)
13.80
(6)
 16.  mahelpet's noname   [normal] 1 0 0 0  (13.86)  3.83
(16)
0.00
(1)
12.17
(16)
85.20
(16)
85.02
(16)
78.47
(13)
72.17
(13)
21.29
(13)
36.59
(13)
34.57
(13)
63.41
(13)
15.70
(16)
31.91
(13)
4.88
(13)
14.13
(15)
33.63
(13)
53.56
(16)
49.11
(16)
50.56
(16)
34.56
(16)
14.87
(14)