Homepage

Description

This dataset consists of features of handwritten numerals (`0'--`9') extracted from a collection of Dutch utility maps. 200 patterns per class (for a total of 2,000 patterns) have been digitized in binary images. These digits are represented in terms of the following six feature sets (files): 1. mfeat-fou: 76 Fourier coefficients of the character shapes; 2. mfeat-fac: 216 profile correlations; 3. mfeat-kar: 64 Karhunen-Love coefficients; 4. mfeat-pix: 240 pixel averages in 2 x 3 windows; 5. mfeat-zer: 47 Zernike moments; 6. mfeat-mor: 6 morphological features. In each file the 2000 patterns are stored in ASCI on 2000 lines. The first 200 patterns are of class `0', followed by sets of 200 patterns for each of the classes `1' - `9'. Corresponding patterns in different feature sets (files) correspond to the same original character. The source image dataset is lost. Using the pixel-dataset (mfeat-pix) sampled versions of the original images may be obtained (15 x 16 pixels).

Related Papers

• Pavel Paclik and Robert P W Duin and Geert M. P. van Kempen and Reinhard Kohlus. On Feature Selection with Measurement Cost and Grouped Features. Pattern Recognition Group, Delft University of Technology. [link]
• Simon Perkins and James Theiler. Online Feature Selection using Grafting. ICML. 2003. [link]
• Jaakko Peltonen and Samuel Kaski. Discriminative Components of Data. IEEE. 2004. [link]
• Xiaoli Z. Fern and Carla Brodley. Cluster Ensembles for High Dimensional Clustering: An Empirical Study. Journal of Machine Learning Research n, a. 2004. [link]
• Xiaofeng He and Partha Niyogi. Locality Preserving Projections. NIPS. 2003. [link]
• [link]