Performance of two multiscale texture algorithms in classifying silver gelatine paper via k-nearest neighbors

Kirsten R. Basinet (1) , Andrew G. Klein (1) , Patrice Abry (2) , Stéphane Roux (2) , Herwig Wendt (3, 4) , Paul Messier (5)

1 WWU - Western Washington University
2 Phys-ENS - Laboratoire de Physique de l'ENS Lyon
3 IRIT-MINDS - CoMputational imagINg anD viSion
4 CNRS - Centre National de la Recherche Scientifique
5 Department of Electrical Engineering, College of Engineering, Qatar University, PO BOX 2713, Doha, Qatar

As part of the Historic Photographic Paper Classification Challenge, a multitude of approaches to quantifying paper texture similarity have been developed. These approaches have yielded encouraging results when applied to very controlled datasets containing photomicrographs of familiar specimens. In this paper, we report on the k-nearest neighbors classification performance of two multiscale analysis-based texture similarity approaches when applied to a much larger reference collection of silver gelatin photographic papers. The clusters for this data set were derived from a visual sorting experiment conducted by art conservators and paper experts later extended through crowd-sourcing. The results show that these texture similarity approaches, when combined with a simple k-nearest neighbors classification algorithm, yield workable performances with accuracy of up to 69%. We discuss this outcome in the context of available data and the cross-validation procedure used, then provide suggestions for improvement.

Mots clés: Texture similarity; Photographic paper; Crowd-sourcing; Multiscale analysis

Domaines: Intelligence artificielle [cs.AI]

Kirsten R. Basinet, Andrew G. Klein, Patrice Abry, Stéphane Roux, Herwig Wendt, et al.. Performance of two multiscale texture algorithms in classifying silver gelatine paper via k-nearest neighbors. IEEE 25th International Conference on Image Processing (ICIP 2018), Oct 2018, Athens, Greece. pp.1028-1032. ⟨hal-02279362⟩