Dataset and subdivisions into train and test sets

Dataset (images)
Classes
Samples x class
Dataset.zip
12
64

Samples x class (training rate)
Train numbers
Test numbers
8 (1/8)
TRAIN_numbers_12_64_8_100
TEST_numbers_12_64_8_100
16 (1/4)
TRAIN_numbers_12_64_16_100 TEST_numbers_12_64_16_100
32 (1/2)
TRAIN_numbers_12_64_32_100 TEST_numbers_12_64_32_100

Pre-computed features (can be used directly for classification)

Colour texture descriptor
File
Dimensions (rows x cols)
Co-occurrence matrices + chromatic features
COOC_and_chroma_features.mat
768 x 9
Co-occurrence matrices + percentiles
COOC_and_colour_percentiles.mat 768 x 14
Gabor features + chromatic features
Gabor_and_chromatic_features.mat 768 x 36
Intra-channel Gabor features
Gabor_features.mat 768 x 96
Integrative co-occurrence matrices
Integrative_COOC.mat 768 x 30
Local BInary Patterns + percentiles LBP_and_colour_percentiles.mat 768 x 45


Labels

Labels can be downloaded here: labels.mat