Digital imaging and signal processing technologies offer opportunities for conservators and affiliated researchers to answer long vexing questions in new and potentially more quantitative ways. This paper will present a collaborative project to systematically and semi-automatically characterize the surface texture of historic photographic papers. Surface texture is a vital attribute defining the appearance of a photographic print. Texture impacts tonal range, rendering of detail, reflectance and conveys subtle, qualitative information about the aesthetic intent of a photographer. During the 20th century, manufacturers created a large diversity of specialized textures. Identification of these textures yields important information about the origin of a photographic print, including the date and the region of manufacture. A texture library of photographic papers containing over 2,000 identified surfaces has been assembled using a simple system for capturing photomicrographs. Lacking a query and retrieval mechanism, this library has only the most basic application for the identification of unknown textures. Addressing this deficit, practical applications are being tested as part of The Museum of Modern Art’s project to characterize photographs from its Thomas Walther collection (funded in part by the Andrew W. Mellon Foundation). Paper texture is being documented by reflectance transformation imaging (RTI) and raking light. RTI and raking light data have been collected on a microscopic scale for dozens of samples, including (1) 80 samples about 30% of which have known matches and (2) 90 samples in 3 sets of 30: (a) 10 texture samples from different locations on one piece of paper, (b) 10 texture samples from different pieces of paper taken from the same manufacturer’s package, and (c) 10 texture samples from the same manufacturer, manufactured to the same specifications and in the same time period but from different manufacturer packages. Using these data, automatic classification procedures have been developed through a collaborative competition by teams at different universities. Each team used a different strategy for deriving the most accurate and efficient algorithm for matching texture images from an unknown sample to a short list of identified references with similar textures gleaned from the library of known textures. The results of this competition will be discussed with a summary of remaining challenges. The classification procedures generally divide into two parts: Feature vector extraction from the images followed by similarity evaluation of the feature vectors. For these tasks many algorithms are plausible with strengths and weaknesses dependent on the peculiarities of materials being analyzed. For example Fourier, wavelet, and multi-fractal analysis may have greater or lesser success on certain types of surfaces based on the physical characteristics, including isotropy and roughness, of the sample. The performance of the better schemes will be summarized in the presentation. The techniques developed through the challenge may have applications for rapidly and inexpensively assembling texture libraries of other photographic papers, such as inkjet papers, and of other materials such as textiles and painted surfaces and for accessing these texture collections through database query and retrieval methods.
Paul Messier is the head of the Lens Media Lab at Yale University's Institute for the Preservation of Cultural Heritage. the LML is devoted to materials-based research on the 20th century photographic print.