Description

A dataset containing pictures of 40 individuals under 10 different lightings. Each image has 92 x 112 pixels. Structured as a 4-tensor with modes 92 x 112 x 40 x 10. The data is now stored in figshare https://ndownloader.figshare.com/files/28005669

Usage

Format

A Tensor object with modes 92 x 112 x 40 x 10. The first two modes correspond to the image pixels, the third mode corresponds to the individual, and the last mode correpsonds to the lighting.

Source

https://www.kaggle.com/kasikrit/att-database-of-faces

References

AT&T Laboratories Cambridge. https://www.kaggle.com/kasikrit/att-database-of-faces

F. Samaria, A. Harter, "Parameterisation of a Stochastic Model for Human Face Identification". IEEE Workshop on Applications of Computer Vision 1994.

See Also

plot_orl