Reconstructing the past appearance of paintings from archival film photographs requires reliable colour correction methods. Such photographs are often subject to exposure variability, illumination differences, film specific colour shifts, and dye instabilities, while frequently lacking embedded colour reference targets. This makes it difficult to assess the accuracy of colour correction and to distinguish genuine material changes artworks from film-induced distortions. Our study addresses this gap by investigating how accurately analogue film photographs can be colour-corrected and how visual change can be quantified, both when reference targets are available and when they are not. For this purpose, a controlled dataset was developed by photographing two mock-up paintings under various lighting and exposure conditions. The films were digitised and paired with hyperspectral scans of the paintings, acquired both before and after the accelerated UVA ageing. Several correction strategies were evaluated, including colour chart-based calibration and methods using stable pigment regions, with root polynomial regression achieving the most consistent and accurate results in terms of the Delta E colour difference relative to the hyperspectral ground truth. Additionally, a reference-free change detection method combining the SAM and RMSE spectral difference metrics was applied to highlight potential pigment colour changes directly from the film photographs, without relying on reference targets or material analysis maps.

Historical lenticular films, such as those created with the Kodacolor process, encode color information in a distinctive spatial format. This structure requires specialized techniques for accurate color reconstruction. While recent signal processing approaches like doLCE and deep learning methods like deep-doLCE have advanced automated color recovery, they often fail with cases such as curved lenticules, low-contrast, or badly captured regions. We propose a human-in-the-loop (HITL) deep learning framework which is designed for color reconstruction in lenticular films. Our approach introduces an editable, vector-based representation of lenticule boundaries, allowing experts to interactively refine boundary positions before color extraction and demosaicing. This decoupled architecture enables targeted corrections and iterative fine-tuning, embedding expert knowledge into the detection model and improving robustness across challenging frames. To preserve image details using information solely present in the original silver emulsion, we merge the reconstructed chrominance with the original film scan's luminance. We evaluate our pipeline on a challenging lenticular film sequence where previous automated approaches fail and the reconstructed colors are not suitable for exhibition. In contrast, our HITL approach successfully produces high-quality, exhibitable color reconstructions with preserved texture. This work is the first to combine expert guidance, editable intermediate representations, and texture-preserving post-processing for lenticular film color reconstruction, advancing the state of the art in this field.

Due to their fragility and uniqueness, valuable works of art are largely unavailable for direct interaction with the public. They are usually displayed in conditions that restrict access, which inevitably limits the visitor experience, particularly for audiences with reduced mobility and visual impairments. Furthermore, related information is not easily conveyed because of the limited time visitors have to browse the museum, and because it is usually presented in jargon that assumes scientific knowledge, or is presented out of context. In this paper, we aim to address all these aspects simultaneously, using the painting The Scream (1910?) from the MUNCH Museum in Oslo as a case study. A combination of high-resolution imaging and computer vision algorithms was used to identify and separate the color surface features for two areas of the painting. The information was then presented to visitors by 3D printing the two areas and placing the color surface features at different heights, creating a relief structure that could be touched and inspected up close. The novel and highly accessible experience was evaluated at the InArt24 conference and at the KHM Museum in Oslo, Norway. During the evaluations, participants reported that interacting with the physical 3D replicas stimulated their interest and motivation, and increased their satisfaction when learning about the painting's color characteristics. These initial findings suggest that presenting information about art objects through multisensory experiences can greatly enhance accessibility and engagement for diverse audiences.

A multimodal optomechatronics system is presented for measuring and monitoring change in cultural heritage objects exposed to environmental condition fluctuations or conservation treatments. It combines structured light, 3D colour digital image correlation and multispectral imaging, delivering information about an object's 3D shape, displacements, strains and reflectivity. The high functionality and applicability of the system are presented with the example of historical parchment subjected to changes in relative humidity.

Observer metamerism (OM) is the name given to the variability between the color matches that individual observers consider accurate. The standard color imaging approach, which uses color-matching functions of a single representative observer, does not accurately represent every individual observer’s perceptual properties. This paper investigates OM in color displays and proposes a quantitative assessment of the OM distribution across the chromaticity diagram. An OM metric is calculated from a database of individual LMS cone fundamentals and the spectral power distributions of the display’s primaries. Additionally, a visualization method is suggested to map the distribution of OM across the display’s color gamut. Through numerical assessment of OM using two distinct publicly available sets of individual observers’ functions, the influence of the selected dataset on the intensity and distribution of OM has been underscored. The case study of digital cinema has been investigated, specifically the transition from xenon-arc to laser projectors. The resulting heatmaps represent the “topography” of OM for both types of projectors. The paper also presents color difference values, showing that achromatic highlights could be particularly prone to disagreements between observers in laser-based cinema theaters. Overall, this study provides valuable resources for display manufacturers and researchers, offering insights into observer metamerism and facilitating the development of improved display technologies.

Imaging sensors are linear over a large part of their operational range. Nevertheless, their behavior becomes non-linear when approaching saturation. This is undesired if such sensors are used for scientific measurements. In this work, a simple and efficient off-chip method is proposed for image sensor linearization. First, the sensor response is characterized with a constant irradiance and a sequence of captures at several integration times. Then a 1D look-up table is calculated to compensate for the nonlinear range. This LUT can be applied to the raw sensor data before further postprocessing. The higher signal-to-noise ratio of captured data is used to demonstrate the benefit of the extended linear range. The proposed method can restore linearity while being easy to implement and computationally efficient.

A material-based approach for the digital restoration of chromogenic photographic and film materials affected by dye fading is proposed. Through a digital reconstruction of the original optical properties, the proposed restoration methodology approximates the original color appearance in a non-subjective manner, thus improving the results compared to conventional RGB tonal re-adjustment of the film scan both in terms of quality and presumed faithfulness to original appearance. In order to do so, the degree of fading is derived from neutral black parts of the film's image content, and the knowledge of the film material's spectral densities is used to digitally reconstruct the colors corresponding to the material's original dye concentrations and render them in an RGB space. For a comparison, results from conventional re-grading were adjusted to render them most similar-and thus comparable-to the results of the proposed spectrally informed digital unfading. The restored images obtained through spectrally informed unfading were deemed clearly superior in terms of color subtlety, color faithfulness and coherence.

The negative-positive chromogenic process, despite being an important milestone in the evolution of color motion picture film technology, exhibits significant fading of its image dyes, leading to the loss of chromatic integrity. A complete spectral approach for the digital restoration of chromogenic film is proposed. A material-based image processing method allows to extract the residual color information associated with the analytical densities, selectively enhance the faded dyes, and finally recreate the original aesthetics by associating the spectral properties of the film stock.

We propose the first accurate digitization and color reconstruction process for historical lenticular film that is robust to artifacts. Lenticular films emerged in the 1920s and were one of the first technologies that permitted to capture full color information in motion. The technology leverages an RGB filter and cylindrical lenticules embossed on the film surface to encode the color in the horizontal spatial dimension of the image. To project the pictures the encoding process was reversed using an appropriate analog device. In this work, we introduce an automated, fully digital pipeline to process the scan of lenticular films and colorize the image. Our method merges deep learning with a model-based approach in order to maximize the performance while making sure that the reconstructed colored images truthfully match the encoded color information. Our model employs different strategies to achieve an effective color reconstruction, in particular (i) we use data augmentation to create a robust lenticule segmentation network, (ii) we fit the lenticules raster prediction to obtain a precise vectorial lenticule localization, and (iii) we train a colorization network that predicts interpolation coefficients in order to obtain a truthful colorization. We validate the proposed method on a lenticular film dataset and compare it to other approaches. Since no colored groundtruth is available as reference, we conduct a user study to validate our method in a subjective manner. The results of the study show that the proposed method is largely preferred with respect to other existing and baseline methods.

The digital reproduction of a historical motion picture should resemble as much as possible the analog film projection at the time of the movie release. Nowadays, practices of capturing digital images of films do not properly consider the fundamental elements and conditions of the original film projection. The typical rigid three-band (RGB) capture cannot adapt to the multitude of historical color film stocks to be digitized, and the diffuse illumination on the film generally used by standard digital scanning devices is unable to guarantee the proper visual rendition of the original analog projection of film prints. In order to overcome these problems, we designed and built a novel multispectral imaging system that illuminates the film with a condensed light beam. The new imaging system and the computational pipeline were tested on an assorted set of photographic colors. The accuracy of the multispectral captures was tested by comparison with corresponding spectrally resolved point-based radiometric measurements of the light reflected by a screen during analog projection. The presented optical design represents an excellent solution for the creation of a new multispectral motion picture scanner prototype. The LED-based illumination system coupled with a film transport mechanism can be the core concept of a promising new generation of motion picture film scanners.

The projection on screen has always been the supreme result of cinematography. Thus the digitization of a motion picture should seek to recreate the visual impression in the cinema. The image-forming particles contained in early color films can generate remarkable differences between a film directly observed with diffuse backlighting and its image projected on screen. In the recent decades this discrepancy has been largely overlooked by the film preservation community (film curators, scanner manufacturers, colorists, etc.). This paper re-establishes the importance of referring to the visual impression in the cinema, and describes the spectral variation of the Callier effect that can significantly alter early film colors when digitized. We have introduced the term “chromatic Callier effect” and described its repercussions on film digitization reporting case studies of tinted and toned film prints. The experimental results highlight that important changes are required in the optical design of film scanners to improve the digitization of motion pictures.

The majority of color film heritage shot between the 1940s and the 1980s is faded. The bleaching of dyes cannot be reversed with chemical methods. Digital technologies can provide the means to recover faded colors. Still, the result of the digital unfading depends on the amount of residual color present in the film, the quality of the image capture operation, and the efficacy of the digital image processing. This paper first discusses the strategies to best capture the residual color information in the film, and then presents a processing method that consistently improves the digital restoration of the most common type of fading: the bleaching of the cyan dye that results in the typical pinkish cast of historical photographs.

The directional arrangement of the illumination plays an important role on image contrast and sharpness of silver-based photographic film. This paper explores how the directional arrangement of light (directed or diffuse) affects cinematographic colors. The experimental results show that a consistent color difference can be observed between the image projected on screen (directed illumination) and the image acquired with a scanner (diffuse illumination), no matter how well the scanner is calibrated. This fact has to be properly considered in film scanning and color correction; otherwise early color films can be notably distorted during the digitization process. Introduction For many decades the public exhibition of motion pictures took place exclusively in cinema theaters. Projection on cinema screens has always been the supreme display of film productions, even when ‘home cinema’ became the most common modality of film consumption. The extreme image magnification of movie projection requires a very high luminous flux on the screen. To increase the luminous flux, in traditional film projectors the light source is located in the focus of a parabolic mirror, which directs the light towards the condenser. The condenser consists of a set of lenses that focus the light further and direct it to the main lens assembly, which images the photographic emulsion on the reflecting cinema screen. Directed illumination is the term that will be used here to describe how a film is illuminated in a projector (Fig. 1-a). To date, almost all cinema theaters around the globe have converted to digital projection, and traditional film projectors are only used by a restricted circle of film enthusiasts. Therefore, the easy access to the content of film reels relies only on their digital replica, which is supposed to create a digital visualization that matches the cinematographic images created by ‘old-style’ film projectors [1]. The vast majority of film scanners create digital replicas of cinematographic images illuminating the film with a light diffuser (e.g. opal glass, or integrating sphere), which provides a diffuse illumination (Fig. 1-b). This type of illumination has the advantage of easily obtaining a uniform illumination across the film gate, and reduces the appearances of blemishes, such as dust and scratches. The directed illumination of film projectors and, on the other hand, the diffuse illumination of film scanners are exact opposites as for the arrangement of ray directions. In directed illumination each point of the film receives light from only one direction, while in diffuse illumination each point of the film receives light from all directions. Some film types illuminated in these two opposite manners create quite different images (e.g. Fig. 1-c and 1-d), and this difference can lead to digital replicas with bad color reproduction. Figure 1. Below: Schemes of directed (a) and diffuse (b) illuminations. Above: Corresponding images of a photographic film (c and d respectively). The film is a print of “Das Cabinet des Dr. Caligari” (R. Wiene, 1920) that underwent metallic toning. This paper highlights the importance of the type of illumination in film digitization, exploring how the directional arrangement of light affects cinematographic colors, providing a contribution to the development of new approaches for the digitization and restoration of film colors. Scientific background The directional arrangement of the rays illuminating a silver-based photographic image plays a fundamental role in its sharpness and contrast [2]. Figure 1 depicts the two different types of illumination at the bottom, and shows the corresponding acquired images at the top. The illustration 1-a depicts the setup in which the film is illuminated by means of a condenser, which provides aligned light rays (directed illumination). Contrarily, in illustration 1-b the film is illuminated by means of a diffuser, which provides scattered light rays (diffuse illumination). The resulting images of a metal-based photographic film adopting the two types of illumination have different sharpness and contrast. In directed illumination the image (Fig. 1-c) is much ‘crisper’ and scratches are emphasized; in diffuse illumination the image (Fig. 1-d) appears ‘softer’, with smoothed details and lower contrast. This discrepancy has been known over a long period of time in black-and-white still-photography, observing the differences between the prints created with condenser or diffuser enlargers. The phenomenon became known as ‘Callier effect’, named after the scientist who defined the Q-factor, i.e. the ratio between the 188 © 2017 Society for Imaging Science and Technology optical densities of a photographic film measured in directed and diffuse illuminations [3]. The Callier effect is determined by the scattering phenomena at the silver particles, which are known to be wavelength-dependent [4]. As a consequence, the images created with directed and diffused illuminations not only differ in sharpness and contrast, but they also have different colors (as the images in Fig. 1). To date, no study has been published on the spectral dependence of the Callier effect (probably due to the fact that this color shift has no influence in the printing of black-and-white photographs). The present work fills this gap by investigating experimentally the relation between the Q-factor and the wavelength of light. The results of this study are particularly interesting for several types of early color films. The images on “modern” color film (i.e. chromogenic monopack) are constituted of dyes that scarcely scatter light; therefore these images do not significantly change with the type of illumination. On the other hand, in case of photographic color images comprised of scattering particles (usually metallic silver) [5], for instance additive screen processes, such as Autochrome and Dufaycolor, or applied colors, such as hand and stencil coloring, as well as tinted and toned films, light is intensely scattered and a strong Callier effect is produced. If this phenomenon is not properly considered, the appearance of applied colors, which were used in early cinema for metaphorical associations and to articulate the narrative structure, risks to be altered significantly by the digitization. In fact, the different illumination manner determines an intrinsic color mismatch, which is not related to the capability of the film scanner to accurately measure colors. Measuring setup The experimental method used here was to acquire two multispectral transmittance images of a photographic film using directed and diffuse illuminations. The measuring setup is depicted in Figure 2. The light was provided by a plasma lamp, coupled with a liquid light guide and a collimating adapter, generating a parallel broadband light beam. An aspheric condenser lens focused the light into a linear variable interference band-pass filter, where the center wavelength of the passed band shifts linearly across its length. Two other aspheric lenses focused the spectrally selected diverging beam coming out of the filter (FWHM = 20 nm) at the principal plane of the camera objective. The imaging system consisted of a 65 mm f/2.8 macro lens and a 16 Megapixel full-frame CCD monochrome camera. Each multispectral image was created combining 31 images between 400 and 700 nm, with a spectral step of 10 nm. The calculation of film transmittances was done with a flat-field correction, referring to the blank images without film and dark images. A light diffuser could be inserted just before the film to be imaged (see Fig. 2), switching from directed to diffuse illumination. Figure 2. The measuring setup adopted for the acquisition of the multispectral images Spectral dependence of the Q-factor The Q-factor (Q) is the curve obtained by plotting the ratio between the optical densities of a photographic film measured in directed (OD∥) and diffuse (OD∦) illuminations as a function of