olivetti_faces: Olivetti Faces dataset

In jlmelville/snedata: SNE Simulation Dataset Functions

Olivetti Faces dataset

Description

SNE benchmarking data.

Usage

olivetti_faces()

Format

A data frame with 400 rows and 4097 variables.

Details

Returns the Olivetti Faces dataset in a data frame reformatted to have one face per row, rather than column. This is a series of 400 images (with dimension 64 x 64) of 40 individual's faces, with ten different poses per person.

The variables are as follows:

• px1, px2, px3 ... px560 8-bit grayscale pixel values (0-255). The pixel index starts at the top right of the image (px1) and are then stored column-wise.

• Label An integer in the range (1-40) indicating the person.

Each row has a name with the format "<face>_<pose>", where <face> is the index of the face, and <pose> is the index of the pose, e.g. the row with name 20_10 is the tenth pose of the twentieth face.

Value

The Olivetti Faces dataset as a dataframe.

Note

Requires the RnavGraphImageData package to be installed and loaded.

See Also

• Sam Roweis' dataset web page: http://www.cs.nyu.edu/~roweis/data.html.

• Each row can be visualized as an image using show_olivetti_face.

jlmelville/snedata documentation built on Jan. 13, 2024, 2:06 a.m.