Skin Tone Results     (training)     (back to index)

Size of test dataset: 28

Percentage of successful automatic results: >95%

Percentage of less successful automatic results: <5%

Comparison with Grabler09: While Grabler09 can successfully transfer the selected region in step 1, it cannot adapt adjustment parameters based on the image features in this selected region. Thus, our modified Grabler09 technique just averages the the parameter values across the demonstrations, which produces results that are identical to the Average results.

Sample of successful results

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09
Sample of less successful results

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09

Input
Macro Result (Automatic)
Ground Truth
Average
Grabler09