CMV/TextureClassification - Department of Computer Science and Engineering

Downloads

Texture classification (32 Brodatz textures)

Description:

Image data for texture classification involving 32 textures from the Brodatz album.

Examples of images in the dataset can be seen below.

Original reference for the image data:

[1] Valkealahti K & Oja E (1998). Reduced multidimensional cooccurrence histograms in texture classification. IEEE Transactions on Pattern Analysis and Machine Intelligence 20:90-94.

Our publications where this image data has been used:

[2] Ojala T, Pietikäinen M & Kyllönen J (1999) Gray level cooccurrence histograms via learning vector quantization. Proc. 11^th Scandinavian Conference on Image Analysis, Kangerlussuaq, Greenland, 103-108. (Abstract) (Full paper)

[3] Ojala T, Valkealahti K, Oja E & Pietikäinen M (2000) Texture discrimination with multidimensional distributions of signed differences. Proc. Fourth Asian Conference on Computer Vision, Taipei, Taiwan, 2:1082-1088. (Abstract) (Full paper)

[4] Ojala T, Valkealahti K, Oja E & Pietikäinen M (2000) Texture discrimination with multidimensional distributions of signed differences. Pattern Recognition, accepted.

Other publications where this image data has been used:

Acknowledgements:

We wish to thank Dr. Kimmo Valkealahti from Nokia Research Center and Professor Erkki Oja from Helsinki University of Technology for providing these images.

Detailed information:

Type of problem:	texture classification
Number of texture classes:	32 textures from the Brodatz album
Number of images (samples):	64 images per class (hence 2048 images in total), which comprise following subsets of images: 16 "original" images, 16 rotated versions of the "original" images, 16 scaled versions of the "original" images, 16 rotated and scaled versions of the "original" images
Image size:	64x64 pixels
Image format:	xv (Khoros Visualization image file)
Pixel format:	8 bit monochrome
Preprocessing:	histogram equalization: each 64x64 sample has uniform gray level histogram

Contents of the zip file:

README	This file.
(texture)_(idx)_(type).xv	An image named as follows:
(texture)	texture class (bark, beachsand, beans, burlap, d10, d11, d4, d51, d52, d5, d6, d95, fieldstone, grass, ice, image09, image15, image17, image19, paper, peb54, pigskin, pressedcl, raffia2, raffia, reptile, ricepaper, seafan, straw2, tree, water, woodgrain)
(idx)	running index (0-15) corresponding to the 16 "original" images
(type)	o original, r rotated, s scaled, rs rotated and scaled
ho10_(trial)_train.txt ho10_(trial)_test.txt	In references [2-4] image data was randomly divided into ten training and testing sets of 1024 images each. These files contain the names of the images used for training and testing in trial (trial) (0-9), respectively.