A4 Conference proceedings

Semisynthetic ground truth for dirt particle counting and classification methods


Publication Details
Authors: Strokina Nataliya, Mankki Aki, Eerola Tuomas, Lensu Lasse, Käyhkö Jari, Kälviäinen Heikki
Editors of book: K. Ikeuchi
Publication year: 2011
Language: English
Related Journal or Series Information: IAPR Conference on Machine Vision Applications
Journal acronym: MVA
ISBN: 978-4-901122-111-5
JUFO-Level of this publication: 0
Open Access: Not an Open Access publication

Abstract
In the evaluation of dirt inclusions in paper, the attention is paid not only to the quantity of dirt but also to the type of dirt particles. Automatic classification methods can be designed for the task, but there should also exist proper evaluation data to truthfully compare the methods. For such comprehensive evaluations, reliable ground truth is essential. To get suitable samples of paper with expert-annotated dirt particles in each can be considered as a too complicated, laborious, and time-consuming task. In the present work, laboratory personnel produced pulp and dirt particles for the samples as pure as possible. Consequently, the initial data was provided as a set of paper sheets with a single dirt type in each. In order to combine dirt particles of different types in one image of a paper sheet and to know the exact location and the type of the particles, a semisynthetic method for generating the ground truth was developed. This paper introduces the algorithm for the purpose. It is shown how the background for dirt particle images is generated, what is the principle of dirt particle placing over the background, and how the problem of color normalization is solved. Finally, the generated background is statistically evaluated, and segmentation result of the dirt particles is used to estimate the quality of the semisynthetic images.

Last updated on 2017-08-06 at 16:57