James R. Mansfield

Near-infrared spectroscopic imaging in art conservation: investigation of drawing constituents

AbstractThe remote-sensing technique of spectroscopic imaging has been adapted to the non-destructive examination of works of art.The principleof near-infrared reflectance spectroscopic imaging is explained, and our instrumentation for art examination described. The technique allowsthe art materials to be distinguished by their composition, and under-drawings revealed. The initial results indicate that even over limitedwavelength ranges (650–1040 nm) and with relatively coarse spectral resolution (10 nm) a number of pigments can be distinguished on thebasis of variations in spectral properties such as spectral slope and the presence or absence of absorption bands. Software adapted from theremote-sensing image-processing field has been used to successfully map areas of different brown and black pigments across a drawing.Non-destructive identification of pigments can be used to address issues of attribution, age dating, and conservation.An additional advantageof this technique is that it can be performed off-site using portable instrumentation, and under relatively benign lighting conditions. Thetechnique has been applied to the examination of a 15th-century drawing,Untitled (The Holy Trinity), in the collection of the Winnipeg ArtGallery. Multivariate image analysis produced a set of principal component (PC) images highlighting different materials’ aspects of thedrawing. A color composite image produced from the PC images provided a direct visualization of the compositional characteristics of thework. Features of the under-drawing have been exposed, and its material tentatively identified as charcoal, by comparison with reference data.© 2003 Editions scientifiques et medicales Elsevier SAS. All rights reserved.

Near infrared spectroscopic reflectance imaging: a new tool in art conservation

The application of infrared spectroscopic imaging to non-destructive examination of works of art is described. Its advantages over infrared photography and reflectography are discussed, in particular its ability to provide spectroscopic information, which potentially allows identification of pigments, binders, and other materials. Near-infrared spectra of a selection of brown and black pigments are presented. Results are given of the application of infrared spectroscopic imaging to two works of art in different media: an ink drawing and an oil painting.

Neural network classification of infrared spectra of control and Alzheimer's diseased tissue.

Artificial neural network classification methods were applied to infrared spectra of histopathologically confirmed Alzheimers diseased and control brain tissue. Principal component analysis was used as a preprocessing technique for some of these artificial neural networks while others were trained using the original spectra. The leave-one-out method was used for cross-validation and linear discriminant analysis was used as a performance benchmark. In the cases where principal components were used, the artificial neural networks consistently outperformed their linear discriminant counterparts; 100% versus 98% correct classifications, respectively, for the two class problem, and 90% versus 81% for a more complex five class problem. Using the original spectra, only one of the three selected artificial neural network architectures (a variation of the back-propagation algorithm using fuzzy encoding) produced results comparable to the best corresponding principal component cases: 98% and 85% correct classifications for the two and five class problems, respectively.

Noninvasive Assessment of Regional and Temporal Variations in Tissue Oxygenation by Near-Infrared Spectroscopy and Imaging

A combination of near-infrared spectroscopy and discrete wavelength near-infrared imaging is used to noninvasively monitor the forearm during periods of restricted blood outflow (venous outflow restriction) and interrupted blood inflow (ischemia). Multivariate analysis of image and spectral data time courses was used to identify highly correlated spectral and regional domains, while fuzzy C-means clustering of image time courses was used to reveal finer regional heterogeneities in the response of stressed tissues. Localized near-infrared spectroscopy was used to investigate the magnitude of the bulk changes in the tissue optical properties and the variation in tissue oxygenation saturation during venous outflow restriction and complete forearm ischemia. The imaging and spectroscopic analyses revealed highly localized regional variations in tissue oxygen saturation during forearm ischemia as compared to the more diffuse and global response of the forearm during venous outflow restriction.

Fuzzy C-means clustering and principal component analysis of time series from near-infrared imaging of forearm ischemia

Fuzzy C-means clustering and principal components analysis were used to analyze a temporal series of near-IR images taken of a human forearm during periods of venous outflow restriction and complete forearm ischemia. The principal component eigen-time course analysis provided no useful information and the principal component eigen-image analysis gave results that correlated poorly with anatomical features. The fuzzy C-means clustering analysis, on the other hand, showed distinct regional differences in the hemodynamic response and scattering properties of the tissue, which correlated well with the anatomical features of the forearm.

Tissue viability by multispectral near infrared imaging: a fuzzy C-means clustering analysis

Clinically, skin color, temperature, and capillary perfusion are used to assess tissue viability following microvascular tissue transfer. However, clinical signs that arise as a consequence of poor perfusion become evident only after several hours of compromised perfusion. This study demonstrates the potential usefulness of optical/infrared multispectral imaging in the prognosis of tissue viability immediately post-surgery. Multispectral images of a skin flap model acquired within 1 h of surgical elevation are analyzed in comparison to the final 72 h clinical outcome with a high degree of correlation. Regional changes in tissue perfusion and oxygenation present immediately following surgery are differentiated using fuzzy clustering and image processing algorithms. These methodologies reduce the intersubject variability inherent in infrared imaging methods such that the changes in perfusion are reproducible and clearly distinguishable across all subjects. Clinically, an early prognostic indicator of viability such as this would allow for a more timely intervention following surgery in the event of compromised microvasculature.

NEAR INFRARED SPECTROSCOPIC REFLECTANCE IMAGING : SUPERVISED VS. UNSUPERVISED ANALYSIS USING AN ART CONSERVATION APPLICATION

Abstract Near-IR spectroscopic imaging was used to analyze the remnants of a work of art, a 16th century drawing, attributed to the School of Pieter Bruegel the Elder, which had been significantly altered during a cleaning attempt. Using a combination of a CCD camera and a liquid crystal tunable filter (LCTF), near-IR spectroscopic images (650–1050 nm) were collected from the drawing and from a test sample composed of four substances with differing near-IR spectra deposited on a whiteboard surface. Both supervised and unsupervised classification methodologies (linear discriminant analysis (LDA) and fuzzy C-means (FCM) clustering, respectively) were used to analyze the data. FCM clustering, in combination with several spectral normalization routines, proved an excellent data exploration method for the test sample. LDA gave consistently clearer results than the FCM methods, but required a priori knowledge of the spectral properties of the sample, provided, in this case, by the FCM analysis. LDA of the spectroscopic image of the work of art revealed clearly and for the first time the location of regions of the drawing where faint traces of ink residue remained.

Near-infrared spectroscopy for dermatological applications

The clinical diagnosis of many dermatological conditions suffers from inadequacies and a histopathological analysis of skin biopsies remains the standard for confirmation of a diagnosis. We suggest that near-infrared (IR) spectroscopy may be a suitable non-invasive, objective tool for characterizing skin conditions. This paper will highlight our in vivo near-IR spectroscopic characterization of skin tumors presented for differential diagnosis of skin cancer. In vivo visible and near-IR spectra (400‐ 2500 nm) were collected from actinic keratoses, basal cell carcinomata, benign nevi, dysplastic nevi, actinic lentigines and seborrheic keratoses by placing a fiber optic probe on the skin. Linear discriminant analysis (LDA) was used to determine whether spectra could be classified according to lesion type and resulted in accuracies of 70‐98% in differentiating benign from pre-malignant or malignant lesions. Near-IR spectroscopy is a promising non-invasive technique for the screening of skin lesions. # 2002 Published by Elsevier Science B.V.

LDA-Guided Search Engine for the Nonsubjective Analysis of Infrared Microscopic Maps

Acquisition of large data sets from human tissues by infrared (IR) microscopy is now routine. However, processing such large data sets, which may contain more than 10 000 spectra, provides an enormous challenge. Overcoming this challenge and developing nonsubjective methods for the analysis of IR microscopic results remain the major hurdle to developing clinically useful applications. A three-step pattern recognition strategy based upon linear discriminant analysis has been developed for use as a search engine for tissue characterization. The three-step strategy includes a genetic algorithm-guided data reduction step, a classification step based upon linear discriminant analysis, and a final step in which the discriminant coefficients are converted into a visually appealing, nonsubjective representation of the distribution of each class throughout the tissue section. The application of this search engine in the characterization of tumor-bearing skin is demonstrated.

Cancer diagnosis by infrared spectroscopy: methodological aspects

IR spectroscopy is proving to be a powerful tool for the study and diagnosis of cancer. The application of IR spectroscopy to the analysis of cultured tumor cells and grading of breast cancer sections is outlined. Potential sources of error in spectral interpretation due to variations in sample histology and artifacts associated with sample storage and preparation are discussed. The application of statistical techniques to assess differences between spectra and to non-subjectively classify spectra is demonstrated.