Assessing Paper Texture Similarity in Matisse Lithographs Using a Triplet Neural Network
Research Mentor(s)
Andrew G. Klein
Description
The last decade has seen growing interest in the use of signal/image processing and machine learning to help answer research questions in cultural heritage, including those in art scholarship and preservation. This work explores the use of a triplet neural network for assessing the similarity of paper textures in a collection of Henri Matisse's lithographs. The available dataset contains digital photomicrographs of papers in the lithograph collection, consisting of four views: two raking light orientations and both sides of the paper. A triplet neural network is first trained to extract features sensitive to anisotropy, and subsequently used to ensure that all papers in the dataset are in the same orientation and side. Another triplet neural network is then used to extract the texture features that are used to assess paper texture similarity. The extraction performs on two different editions of the dataset. These results can then be used by art conservators and historians to answer questions of art historical significance, such as artist intent.
Document Type
Event
Start Date
May 2022
End Date
May 2022
Location
Carver Gym (Bellingham, Wash.)
Department
CSE -Engineering and Design
Genre/Form
student projects; posters
Type
Image
Rights
Copying of this document in whole or in part is allowable only for scholarly purposes. It is understood, however, that any copying or publication of this document for commercial purposes, or for financial gain, shall not be allowed without the author’s written permission.
Language
English
Format
application/pdf
Assessing Paper Texture Similarity in Matisse Lithographs Using a Triplet Neural Network
Carver Gym (Bellingham, Wash.)
The last decade has seen growing interest in the use of signal/image processing and machine learning to help answer research questions in cultural heritage, including those in art scholarship and preservation. This work explores the use of a triplet neural network for assessing the similarity of paper textures in a collection of Henri Matisse's lithographs. The available dataset contains digital photomicrographs of papers in the lithograph collection, consisting of four views: two raking light orientations and both sides of the paper. A triplet neural network is first trained to extract features sensitive to anisotropy, and subsequently used to ensure that all papers in the dataset are in the same orientation and side. Another triplet neural network is then used to extract the texture features that are used to assess paper texture similarity. The extraction performs on two different editions of the dataset. These results can then be used by art conservators and historians to answer questions of art historical significance, such as artist intent.