Tetsuro Morimoto

BRDF Estimation of Structural Color Object by Using Hyper Spectral Image

In computer graphics, there are several methods to render thin film interference effects, but it sets the parameters of interference manually. To reconstruct more accurate appearance of the thin film, we need to estimate the BRDF parameters of the thin film, the refractive index and film thickness. In this paper, we propose a novel method for estimating the BRDF of the thin film interference and representing the appearance accurately. Using peak wavelengths which enhance the spectral intensities, we estimate the refractive index and film thickness separately, and then it reduces the computational time of estimation comparied with the steepest descent method. In the experiments, we evaluate the accuracy of estimated parameters and reconstructed BRDF to show the effectiveness of our method.

Multispectral imaging for material analysis in an outdoor environment using Normalized Cuts

Multispectral imaging has promising and wide application areas, including medical imaging, remote sensing, and cultural asset preservation. In particular, cultural assets are often damaged by microorganisms such as moss and mold. Thus, asset preservation requires measuring the kinds and extent of microorganisms by obtaining spectral information in wide areas. This process requires developing an efficient spectral sensing system that can obtain data for wide areas as well as segmentation methods to identify those locations. This paper describes a new multispectral imaging system applicable to wide areas. Our design allows the system to have a wide field of view of high resolution with low noise and negligible distortion. We can apply this system to measuring the surface spectrum on an object surface in an outdoor environment. For determining the distribution of microorganisms, we developed a multispectral image segmentation method using the data obtained by our system. Finally, we applied our system and segmentation method to the data from the bas-relief of the Bayon Temple in the Angkor ruin, and we identified the classes and distribution areas of the microorganisms.

Reconstructing Shape and Appearance of Thin Film Objects with Hyper Spectral Sensor

Modeling the shape and appearance of a thin film object has promising applications such as heritage-modeling and industrial inspections. In the same time, such modeling is a frontier of computer vision and contains various challenging issues to be solved. In particular, thin film colors show iridescence along the view and lighting directions and how to acquire and formalize the spectral iridescence for shape estimation. This paper aims to model the shapes and appearances of thin film objects from measured reflectance spectra. Thin film reflectance is represented by the incident angle on the object surface, the refractive index and the film thickness. First, we estimate the incident angle of a surface patch on a thin film based on monotonically increasing peak intensities. Then, we apply a characteristics strip expansion method to the peak intensity for estimating the surface normal of the patch. Based on this shape estimation, we estimate refractive index and film thickness from iridescence variance. We experimentally evaluate the accuracy of the estimated shape and estimated parameters. We also demonstrate to reconstruct appearances based on the shape and parameters.

Reconstructing Shapes and Appearances of Thin Film Objects Using RGB Images

Reconstruction of shapes and appearances of thin film objects can be applied to many fields such as industrial inspection, biological analysis, and archaeologic research. However, it comes with many challenging issues because the appearances of thin film can change dramatically depending on view and light directions. The appearance is deeply dependent on not only the shapes but also the optical parameters of thin film. In this paper, we propose a novel method to estimate shapes and film thickness. First, we narrow down candidates of zenith angle by degree of polarization and determine it by the intensity of thin film which increases monotonically along the zenith angle. Second, we determine azimuth angle from occluding boundaries. Finally, we estimate the film thickness by comparing a look-up table of color along the thickness and zenith angle with captured images. We experimentally evaluated the accuracy of estimated shapes and appearances and found that our proposed method is effective.

Accuracy of the spider model in decomposing layered surfaces

The surface of most natural objects is composed of two or more layers whose optical properties jointly determine the surfaces overall reflectance. Light transmission through these layers can be approximated by using the Lambert-Beer (LB) model, which provides a good trade-off between the accuracy and simplicity to handle layer decomposition. Recently, a layer decomposition based on the LB-based model is proposed. Assuming surfaces with two layers, it estimates the reflectance of top and bottom layers, as well as the opacity of the top layer. The method introduces the “spider model”, which is named after the color distribution in the RGB space that resembles the shape of spiders. In this paper, we intend to verify the accuracy of the spider model and the optical model where it is based on (i.e., the LB-based model). We verify the LB-based model by comparing to the Kubelka-Munk (KM) model, which has previously been shown to be reliably accurate. The benefits of layer decomposition are easy to notice. First, many computer vision algorithms assume a single layer, and tend to fail when encountering multi-layered surfaces. Second, knowing the optical properties of each layer can provide further knowledge of the target objects.