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] CLresult byhuangyuan's unet (deepv)
set \ criterionCSOSUSMENEOCCACOCCIIIEAMSRMCIGCELCEBCEGBCEBGMSCSSCVDLAVINVINMIMARIJCDCFMIWIWIINBDE
#3 - all99.940.000.000.000.0099.9899.980.020.040.0499.9635.130.04-31.833.340.040.130.110.140.1399.9499.8799.870.0699.940.090.5099.500.0999.8099.7499.8799.8799.8799.870.04
#4 - all99.500.000.000.000.0099.8899.920.060.130.1399.8728.610.13-37.378.250.130.990.931.060.9999.5099.0199.010.5099.470.572.5597.400.4998.7498.1699.0799.0799.1099.040.26
#5 - all99.380.000.000.000.0099.9399.950.060.130.1399.8713.680.13-37.2012.150.131.240.971.501.2499.3898.7798.760.6299.100.652.5397.260.4498.8398.2799.1399.1399.1999.070.48
#6 - all99.100.000.000.000.0099.9399.970.040.070.2699.9313.750.08-44.8720.770.121.791.552.031.7999.1098.2198.210.9099.100.913.1596.390.9197.5496.3598.1498.1498.2098.080.99
#7 - all87.840.000.0011.688.64100.00100.000.060.090.2999.9114.600.13-45.559.800.155.913.678.896.6496.4293.6993.143.5895.772.116.7692.761.9093.4389.7394.5994.6093.2995.932.85
#8 - all99.230.000.000.000.00100.00100.000.060.150.1199.8514.080.12-42.565.820.121.521.391.651.5299.2398.4898.480.7799.090.822.4697.440.3998.4597.3898.6798.6798.6798.670.24
#9 - all54.940.000.0043.5044.2099.8199.926.558.8313.9091.1710.0310.74-32.609.4511.0520.4513.1430.3123.0078.3472.0071.8816.7777.476.3618.0579.628.6068.9358.8074.0674.2579.8369.0610.37
#10 - all97.710.000.000.000.0099.9599.930.540.891.2899.1110.321.04-42.694.561.064.333.165.584.4197.7195.5795.622.2997.291.915.1894.670.8595.7792.7796.2596.2596.5895.920.68
#11 - all98.550.000.000.000.00100.0099.990.090.130.2999.877.910.18-39.786.330.202.742.323.242.8298.5597.2797.201.4597.711.313.4096.430.4597.7696.1098.0198.0197.8998.130.55
#12 - all90.007.3112.150.000.00100.00100.000.020.040.0699.967.130.05-44.899.460.055.122.618.976.4595.0692.7492.713.3589.072.225.5893.841.1894.6190.9995.2895.2895.5794.994.02
#6 - bark98.630.000.000.000.0099.9999.980.120.350.1999.6512.240.24-45.9115.270.252.691.983.412.7098.6397.3097.311.3798.081.264.4095.180.9297.3195.8997.9097.9098.1497.661.54
#6 - flowers97.520.000.000.000.0099.8899.880.080.170.2299.8312.850.17-41.4413.150.184.282.835.984.5397.5295.4495.312.4896.601.555.3994.221.2696.1794.1096.9696.9696.4297.512.47
#6 - glass98.750.000.000.000.0099.7999.600.150.270.5099.7316.020.31-40.7513.900.342.462.082.862.4898.7597.5597.541.2598.261.244.3395.350.8697.3895.9397.9297.9297.6798.176.69
#6 - man-made99.340.000.000.000.00100.00100.000.040.070.0899.9319.570.07-35.515.040.071.311.031.591.3199.3498.6998.690.6699.270.712.4797.370.4898.5197.6698.8298.8298.9198.720.29
#6 - nature99.400.000.000.000.00100.00100.000.020.030.0499.9713.490.03-20.9022.270.031.200.941.461.2099.4098.8098.800.6099.110.652.2796.810.6698.6198.3199.1599.1599.3998.910.27
#6 - plants99.320.000.000.000.0099.9999.990.040.080.1099.9216.960.09-35.834.540.091.351.211.481.3599.3298.6698.660.6899.290.752.6097.320.4798.4597.5198.7498.7498.6898.800.30
#6 - rock98.340.000.000.000.0099.9399.960.160.260.4499.7419.370.32-41.108.930.343.232.394.093.2598.3496.7496.751.6698.371.424.9594.801.3195.8393.4996.6496.6497.0896.190.74
#6 - stone98.560.000.000.000.00100.00100.000.020.040.1299.9610.780.04-47.5929.300.062.641.803.502.6698.5697.2897.281.4495.861.033.5995.050.5998.6998.2999.1499.1499.3198.970.83
#6 - textile99.720.000.000.000.0099.9399.940.050.080.2099.9219.780.10-40.2310.980.120.550.480.620.5599.7299.4599.450.2899.590.341.1798.730.1999.4299.0999.5499.5499.5399.550.11
#6 - wood88.737.0240.050.000.00100.00100.000.000.000.00100.0012.680.00-39.4320.590.004.762.1210.637.9894.0392.8790.373.6381.902.117.3491.443.1091.7988.4493.8693.9789.6498.502.07
mean95.220.722.612.762.6499.9599.950.410.590.9299.4115.450.70-39.4011.700.733.432.344.953.8597.3495.9295.752.2296.021.404.4395.081.2695.8093.8596.5996.6096.6596.591.79
criterionGraph
correct
over
under
missed
noise
F-measure
imageTexture mosaic (undegraded)Ground truthSegmentation
#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