Note:
See THIS PAGE for more recent experiments where local shift leads to cleaner templates.
(1)
data source: ETHZ shape dataset
Number of training images = 16; Number of elements = 30;
Length of Gabor = 17 pixels; Local normalization or not = 1;
Range of displacement = 4 pixels; Subsample rate = 1 pixel;
Image height and width = 75 by 110 pixels
(2)
data source: ETHZ shape dataset
Number of training images = 35; Number of elements = 30;
Length of Gabor = 17 pixels; Local normalization or not = 1;
Range of displacement = 4 pixels; Subsample rate = 1 pixel;
Image height and width = 105 by 109 pixels
(3)
data source: ETHZ shape dataset
Number of training images = 27; Number of elements = 30;
Length of Gabor = 17 pixels; Local normalization or not = 1;
Range of displacement = 4 pixels; Subsample rate = 1 pixel;
Image height and width = 150 by 59 pixels
(4)
data source: ETHZ shape dataset
Number of training images = 43; Number of elements = 35;
Length of Gabor = 17 pixels; Local normalization or not = 1;
Range of displacement = 4 pixels; Subsample rate = 1 pixel;
Image height and width = 135 by 116 pixels
(5)
data source: ETHZ shape dataset
Number of training images = 44; Number of elements = 40;
Length of Gabor = 25 pixels; Local normalization or not = 1;
Range of displacement = 4 pixels; Subsample rate = 1 pixel;
Image height and width = 165 by 143 pixels