Background

This document provides a brief Users' Guide to the KTH-TIPS image database (KTH is the abbreviation of our university, and TIPS stands for Textures under varying Illumination, Pose and Scale). The guide describes which materials are contained in the database (Section 2), how images were acquired (Section 3) and subsequently cropped to remove the background (Section 4), and we also discuss some non-ideal artifacts, like poor focus, in some pictures (Section 5). This document concludes by outlining how we intend to extend the database in the future (Section 6).

The objectives with this database were to supplement the CUReT image database [1] in two directions, both of which concern extending material classification algorithms to function in real-world conditions. While the CUReT database contains images of 61 materials (over varying pose and illumination, but at constant viewing distance), the aims with KTH-TIPS were:

1. to provide variations in scale as well as variations in pose and illumination. This allows a systematic study of how important unknown viewing distance is to material classification, and provides data for evaluating algorithms intended to be robust to such variations;
2. to provide images of other samples of a subset of the CUReT materials, taken under different settings. We wanted to see whether it would be possible to actually classify materials in the real-world, as opposed to recognising exemplars of materials within a single database.

The cropped database is freely available on the internet [2]. Those interested in the full-size images should contact Eric Hayman (hayman@nada.kth.se).

The database was first presented and used in [3].