A pigment analysis tool for hyperspectral images of cultural heritage artifacts

Abstract

The Gough Map, in the collection at the Bodleian Library, Oxford University, is one of the earliest surviving maps of Britain. Previous research deemed that it was likely created over the 15th century and afterwards it was extensively revised more than once. In 2015, the Gough Map was imaged using a hyperspectral imaging system at the Bodleian Library. The collection of the hyperspectral image (HSI) data was aimed at faded text enhancement for reading and pigment analysis for the material diversity of its composition and potentially the timeline of its creation. In this research, we introduce several methods to analyze the green pigments in the Gough Map, especially the number and spatial distribution of distinct green pigments. One approach, called the Gram Matrix, has been used to estimate the material diversity in a scene (i.e., endmember selection and dimensionality estimation). Here, we use the Gram Matrix technique to study the within-material differences of pigments in the Gough map with common visual color. We develop a pigment analysis tool that extracts visually common pixels, green pigments in this case, from the Gough Map and estimates its material diversity. It reveals that the Gough Map consists of at least six kinds of dominant green pigments. Both historical geographers and cartographic historians will benefit from this work to analyze the pigment diversity using HSI of cultural heritage artifacts.

Stochastic Neighbor Embedding Feature-Based Hyperspectral Image Classification Using 3D Convolutional Neural Network

“…In the Equation ( 4), 2ζ + 1 is the kernel depth along a spectral dimension, and other parameters are the same as in Equation (5).…”

“…HSI is usually used for satellite-image and land-cover analysis [1]. Now it is also used in various applications such as medical image analysis and diagnosis [2], food quality identification [3], material property detection [4], cultural heritage digitalization [5], forensic analysis [6], forest analysis [7], etc. [8][9][10].…”

Neural Networks for Hyperspectral Imaging of Historical Paintings: A Practical Review

“…A range of cutting-edge data processing techniques has particularly been applied in the CH domain to reduce the dimensionality of the dataset, classify and unmix spectral signature to map paint components. Thus, numerous approaches were developed, starting from the conventional multivariate analyses and statistical methods (e.g., spectral angle mapper (SAM) [10][11][12][13], fully constrained least square (FCLS) [14][15][16], principal component analyses (PCA) [17][18][19][20], minimum noise fraction transform (MNF) [21,22], and k-means clustering [23][24][25]), to the more advanced machine learning algorithms (support vector machine (SVM) [26,27], hierarchical clustering [28], embedding techniques [29][30][31][32], MaxD [11,24,33], dictionary learning [34,35]) with a growing interest for neural network algorithms (NNs) [36]. NN-based models first gained a tremendous rise in digital image classification due to their superior ability in feature extraction and pattern recognition [37][38][39][40].…”

Nondestructive Evaluation of Heritage Object Coatings with Four Hyperspectral Imaging Systems

“…However, large training datasets are needed to successfully explore convolutional neural networks (CNNs). New methods for acquiring high-resolution hyperspectral images [25][26][27] or new algorithms for post-processing spectral data [22,[28][29][30][31][32][33] are currently of interest to engineers and researchers working in this field. Similarly, multisensor systems that allow automatic characterisation routines of art objects with high accuracy and precision are desired [34].…”