During my Ph.D. studies at Uppsala University, I acquired and published three texture datasets, I have had requests on hosting these datasets on additional locations other than the university server (address http://cb.uu.se/~gustaf/). Here is an overview: link
The dataset is based on transmission electron microscopy (TEM) images of 15 virus types. The texture samples have been automatically extracted from objects segmented using the method described in . The virus texture dataset was first used in . There is no garatee that the classes are fully discriminable. Here follows a short list of dataset properties:
- 15 texture classes, see Figure 1.
- 100 unique texture patches per class.
- Texture patch size: 41×41 pixels.
- File format: Lossless compressed 16 bit PNG.
- Files are named as follows: class-003-sample-036.tif, where class-003 is the class number and sample-036 is the sample number.
- Virus texture dataset (16bit) as .zip (~2.7 MB) or .7z (~2.3 MB). (updated 2013-01-22)
- Virus texture dataset, intensities resampled to 8bit, as .zip (~2.2 MB) or .7z (~1.8 MB). (The version online before 2013-01-22)
- Text file with class names corresponding to the class numbers used in the image file names: classNames.csv.
- Indices used in the 10-fold cross-validation procedure in manuscript:
- Binary mask:
- Kylberg G., Uppström M., Borgefors G., and Sintorn I.-M. Segmentation of Virus Particle Candidates in Transmission Electron Microscopy Images Journal of Microscopy, 2012, 245: 140-147.
- Kylberg G., Uppström M., and Sintorn I.-M. Virus Texture Analysis Using Local Binary Patterns and Radial Density Profiles In Proceedings of the 16th Iberoamerican Congress on Pattern Recognition (CIARP), LNCS-7042, pp. 573-580, Pucón, Chile, November 2011.