Automatic data-driven design and 3D printing of custom ocular prostheses.

Millions of people require custom ocular prostheses due to eye loss or congenital defects. The current fully manual manufacturing processes used by highly skilled ocularists are time-consuming with varying quality. Additive manufacturing technology has the potential to simplify the manufacture of ocular prosthetics, but existing approaches just replace to various degrees craftsmanship by manual digital design and still require substantial expertise and time. Here we present an automatic digital end-to-end process for producing custom ocular prostheses that uses image data from an anterior segment optical coherence tomography device and considers both shape and appearance. Our approach uses a statistical shape model to predict, based on incomplete surface information of the eye socket, a best fitting prosthesis shape. We use a colour characterized image of the healthy fellow eye to determine and procedurally generate the prosthesis's appearance that matches the fellow eye. The prosthesis is manufactured using a multi-material full-colour 3D printer and postprocessed to satisfy regulatory compliance. We demonstrate the effectiveness of our approach by presenting results for 10 clinic patients who received a 3D printed prosthesis. Compared to a current manual process, our approach requires five times less labour of the ocularist and produces reproducible output.

Cuttlefish: Pushing the Limits of Graphical 3-D Printing.

This article presents Cuttlefish, a 3-D printer driver developed at the Fraunhofer Institute for Computer Graphics Research IGD. Cuttlefish maximizes the reproduction quality of shape and appearance in multimaterial 3-D printing systems. It controls various printing systems and material sets, including color, transparent, and flexible materials. The article provides an overview of Cuttlefish's graphical 3-D printing pipeline and focuses on its unique features: Displaced signed distance fields for robust voxelization of objects not designed for 3-D printing, joint color and translucency reproduction, accurate color and translucency characterization, shape dithering to eliminate quantization-based staircase artifacts, and support for displacement maps and constructive solid geometry operations. The article concludes by showcasing applications of Cuttlefish in printing replacement faces for stop-motion animation, large-scale figurine production, and prosthetic eye printing.

Multi-printer learning framework for efficient optical printer characterization.

A high prediction accuracy of optical printer models is a prerequisite for accurately reproducing visual attributes (color, gloss, translucency) in multimaterial 3D printing. Recently, deep-learning-based models have been proposed, requiring only a moderate number of printed and measured training samples to reach a very high prediction accuracy. In this paper, we present a multi-printer deep learning (MPDL) framework that further improves data efficiency utilizing supporting data from other printers. Experiments on eight multi-material 3D printers demonstrate that the proposed framework can significantly reduce the number of training samples thus the overall printing and measurement efforts. This makes it economically feasible to frequently characterize 3D printers to achieve a high optical reproduction accuracy consistent across different printers and over time, which is crucial for color- and translucency-critical applications.

Inducing robustness and plausibility in deep learning optical 3D printer models.

Optical 3D printer models characterize multimaterial 3D printers by predicting optical or visual quantities from material arrangements or tonal values. Their accuracy and robustness to noisy training data are crucial for 3D printed appearance reproduction. In our recent paper [Opt. Express29, 615 (2021)10.1364/OE.410796], we have proposed a pure deep learning (PDL) optical model and a training strategy achieving high accuracy with a moderate number of training samples. Since the PDL model is essentially a black-box without considering any physical grounding, it is sensitive to outliers or noise of the training data and tends to create physically-implausible tonal-to-optical relationships. In this paper, we propose a methodology to narrow down the degrees-of-freedom of deep-learning based optical printer models by inducing physically plausible constraints and smoothness. Our methodology does not need any additional printed samples for training. We use this approach to introduce the robust plausible deep learning (RPDL) optical printer model enhancing robustness to erroneous and noisy training data as well as physical plausibility of the PDL model for selected tonal-to-optical monotonicity relationships. Our experiments on four state-of-the-art multimaterial 3D printers show that the RPDL model not only almost always corrects implausible tonal-to-optical relationships, but also ensures significantly smoother predictions, without sacrificing accuracy. On small training data, it even outperforms the PDL model in accuracy by up to 8% indicating a better generalization ability.

Perceptually Optimizing Color Look-up Tables.

The quality of ICC profiles with embedded look-up tables (LUTs) depends on multiple factors: 1. the accuracy of the optical printer model, 2. the exploitation of the available gamut combined with the quality of the gamut mapping approach encoded in the B2A-LUTs (backwards LUTs) and 3. the tonal smoothness as well color accuracy of the backwards LUTs. It can be shown that optimizing the smoothness of the LUTs comes at the expense of color accuracy and requires gamut reduction because of internal tonal edges. We present a method to optimize backwards LUTs of existing ICC profiles w.r.t accuracy, smoothness, gamut exploitation and mapping, which can be extended beyond color, e.g. to joint color and translucency backward LUTs. The approach is based on a perceptual difference metric that is used to optimize the LUT's tonal smoothness constrained to preserve both the accuracy of and the relationship between colors.

Deep learning models for optically characterizing 3D printers.

Multi-material 3D printers are able to create material arrangements possessing various optical properties. To reproduce these properties, an optical printer model that accurately predicts optical properties from the printer's control values (tonals) is crucial. We present two deep learning-based models and training strategies for optically characterizing 3D printers that achieve both high accuracy with a moderate number of required training samples. The first one is a Pure Deep Learning (PDL) model that is essentially a black-box without any physical ground and the second one is a Deep-Learning-Linearized Cellular Neugebauer (DLLCN) model that uses deep-learning to multidimensionally linearize the tonal-value-space of a cellular Neugebauer model. We test the models on two six-material polyjetting 3D printers to predict both reflectances and translucency. Results show that both models can achieve accuracies sufficient for most applications with much fewer training prints compared to a regular cellular Neugebauer model.

Characterization and Decomposition of the Natural van der Waals SnSb2Te4 under Compression.

High pressure X-ray diffraction, Raman scattering, and electrical measurements, together with theoretical calculations, which include the analysis of the topological electron density and electronic localization function, evidence the presence of an isostructural phase transition around 2 GPa, a Fermi resonance around 3.5 GPa, and a pressure-induced decomposition of SnSb2Te4 into the high-pressure phases of its parent binary compounds (α-Sb2Te3 and SnTe) above 7 GPa. The internal polyhedral compressibility, the behavior of the Raman-active modes, the electrical behavior, and the nature of its different bonds under compression have been discussed and compared with their parent binary compounds and with related ternary materials. In this context, the Raman spectrum of SnSb2Te4 exhibits vibrational modes that are associated but forbidden in rocksalt-type SnTe; thus showing a novel way to experimentally observe the forbidden vibrational modes of some compounds. Here, some of the bonds are identified with metavalent bonding, which were already observed in their parent binary compounds. The behavior of SnSb2Te4 is framed within the extended orbital radii map of BA2Te4 compounds, so our results pave the way to understand the pressure behavior and stability ranges of other "natural van der Waals" compounds with similar stoichiometry.

Identification of the strong Brønsted acid site in a metal-organic framework solid acid catalyst.

It remains difficult to understand the surface of solid acid catalysts at the molecular level, despite their importance for industrial catalytic applications. A sulfated zirconium-based metal-organic framework, MOF-808-SO4, was previously shown to be a strong solid Brønsted acid material. In this report, we probe the origin of its acidity through an array of spectroscopic, crystallographic and computational characterization techniques. The strongest Brønsted acid site is shown to consist of a specific arrangement of adsorbed water and sulfate moieties on the zirconium clusters. When a water molecule adsorbs to one zirconium atom, it participates in a hydrogen bond with a sulfate moiety that is chelated to a neighbouring zirconium atom; this motif, in turn, results in the presence of a strongly acidic proton. On dehydration, the material loses its acidity. The hydrated sulfated MOF exhibits a good catalytic performance for the dimerization of isobutene (2-methyl-1-propene), and achieves a 100% selectivity for C8 products with a good conversion efficiency.

Cornucopia of Structures in the Pseudobinary System (SnSe) xBi2Se3: A Crystal-Chemical Copycat.

Pseudobinary phases (SnSe) xBi2Se3 exhibit a very diverse structural chemistry characterized by different building blocks, all of which are cutouts of the NaCl type. For SnSe contents between x = 5 and x = 0.5, several new phases were discovered. Next to, for example, Sn4Bi2Se7 ( x = 4) in the NaCl structure type and SnBi4Se7 ( x = 0.5) in the layered defect GeSb2Te4 structure type, there are at least four compounds (0.8 ≤ x ≤ 3) with lillianite-like structures built up from distorted NaCl-type slabs (L4,4-type Sn2.22Bi2.52Se6, L4,5-type Sn9.52Bi10.96Se26, L4,7-type Sn11.49Bi12.39Se30, and L7,7-type Sn3.6Bi3.6Se9). For two of them (L4,7 and L7,7), the cation distributions were determined by resonant X-ray scattering, which also confirmed the presence of significant amounts of cation vacancies. Thermoelectric figures of merit ZT range from 0.04 for Sn4Bi2Se7 to 0.2 for layered SnBi4Se7; this is similar to that of the related compounds SnBi2Te4 or PbBi2Te4. Compounds of the lillianite series exhibit rather low thermal conductivities (∼0.75 W/mK for maximal ZT). More than other "sulfosalts", compounds in the pseudobinary system SnSe-Bi2Se3 adapt to changes in the cation-anion ratio by copying structure types of compounds containing lighter or heavier homologues of Sn, Bi, or Se and can incorporate significant amounts of vacancies. Thus, (SnSe) xBi2Se3 is a multipurpose model system with vast possibilities for substitutional and structural modification aiming at the optimization of thermoelectric or other properties.

Two Principles of Reticular Chemistry Uncovered in a Metal-Organic Framework of Heterotritopic Linkers and Infinite Secondary Building Units.

Structural diversity of metal-organic frameworks (MOFs) has been largely limited to linkers with at most two different types of coordinating groups. MOFs constructed from linkers with three or more nonidentical coordinating groups have not been explored. Here, we report a robust and porous crystalline MOF, Zn3(PBSP)2 or MOF-910, constructed from a novel linker PBSP (phenylyne-1-benzoate, 3-benzosemiquinonate, 5-oxidopyridine) bearing three distinct types of coordinative functionality. The MOF adopts a complex and previously unreported topology termed tto. Our study suggests that simple, symmetric linkers are not a necessity for formation of crystalline extended structures and that new, more complex topologies are attainable with irregular, heterotopic linkers. This work illustrates two principles of reticular chemistry: first, selectivity for helical over straight rod secondary building units (SBUs) is achievable with polyheterotopic linkers, and second, the pitch of the resulting helical SBUs may be fine-tuned based on the metrics of the polyheterotopic linker.

How psychophysical methods influence optimizations of color difference formulas.

For developing color difference formulas, there are several choices to be made on the psychophysical method used for gathering visual (observer) data. We tested three different psychophysical methods: gray scales, constant stimuli, and two-alternative forced choice (2AFC). Our results show that when using gray scales or constant stimuli, assessments of color differences are biased toward lightness differences. This bias is particularly strong in LCD monitor experiments, and also present when using physical paint samples. No such bias is found when using 2AFC. In that case, however, observer responses are affected by other factors that are not accounted for by current color difference formulas. For accurate prediction of relative color differences, our results show, in agreement with other works, that modern color difference formulas do not perform well. We also investigated if the use of digital images as presented on LCD displays is a good alternative to using physical samples. Our results indicate that there are systematic differences between these two media.

How color difference formulas depend on reference pairs in the underlying constant stimuli experiment.

For calculating color differences, the CIEDE2000 and CIE94 equations are widely used and recommended. These equations were derived more than a decade ago, based for a large part on the RIT-Dupont set of visual data. This data was collected from a series of psychophysical tests that use the method of constant stimuli. In this method, observers need to compare the color difference within a sample pair to that between a reference pair. In the current investigation, we show that the color difference equation significantly changes if reference pairs are chosen in the underlying visual experiments that differ from what was used when creating the RIT-Dupont dataset. The investigation is done using metallic paint samples representing two color centers, red and yellow-green. We show that the reproducibility differs for three different reference pairs, and that for modeling the visual data for the yellow-green color center, extra model terms are required as compared to the CIEDE2000 equation. Our results suggest that observers differ in their ability to mentally convert a color difference recognized in a sample pair into an equivalent color difference along the color difference direction represented by the reference pair. We also find that in these tests the tolerance to lightness differences is widened by a factor of 1.3 to 1.6, and that for the red color center the tolerance ellipsoid is rotated by 30° as compared to the CIEDE2000 equation. The latter observations are possibly due to the metallic texture in the samples used for the current experiment.

Measuring isotropic subsurface light transport.

Subsurface light transport can affect the visual appearance of materials significantly. Measuring and modeling this phenomenon is crucial for accurately reproducing colors in printing or for rendering translucent objects on displays. In this paper, we propose an apparatus to measure subsurface light transport employing a reference material to cancel out adverse signals that may bias the results. In contrast to other approaches, the setup enables improved focusing on rough surfaces (e.g. uncoated paper). We derive a measurement equation that may be used to deduce the point spread function (PSF) of subsurface light transport. Main contributions are the usage of spectrally-narrowband exchangeable LEDs allowing spectrally-resolved measurements and an approach based on quadratic programming for reconstructing PSFs in the case of isotropic light transport.

Image-difference prediction: from color to spectral.

We propose a new strategy to evaluate the quality of multi and hyperspectral images, from the perspective of human perception. We define the spectral image difference as the overall perceived difference between two spectral images under a set of specified viewing conditions (illuminants). First, we analyze the stability of seven image-difference features across illuminants, by means of an information-theoretic strategy. We demonstrate, in particular, that in the case of common spectral distortions (spectral gamut mapping, spectral compression, spectral reconstruction), chromatic features vary much more than achromatic ones despite considering chromatic adaptation. Then, we propose two computationally efficient spectral image difference metrics and compare them to the results of a subjective visual experiment. A significant improvement is shown over existing metrics such as the widely used root-mean square error.

Layered germanium tin antimony tellurides: element distribution, nanostructures and thermoelectric properties.

In the system Ge-Sn-Sb-Te, there is a complete solid solution series between GeSb2Te4 and SnSb2Te4. As Sn2Sb2Te5 does not exist, Sn can only partially replace Ge in Ge2Sb2Te5; samples with 75% or more Sn are not homogeneous. The joint refinement of high-resolution synchrotron data measured at the K-absorption edges of Sn, Sb and Te combined with data measured at off-edge wavelengths unambiguously yields the element distribution in 21R-Ge(0.6)Sn(0.4)Sb2Te4 and 9P-Ge(1.3)Sn(0.7)Sb2Te5. In both cases, Sb predominantly concentrates on the position near the van der Waals gaps between distorted rocksalt-type slabs whereas Ge prefers the position in the middle of the slabs. No significant antisite disorder is present. Comparable trends can be found in related compounds; they are due to the single-side coordination of the Te atoms at the van der Waals gap, which can be compensated more effectively by Sb(3+) due to its higher charge in comparison to Ge(2+). The structure model of 21R-Ge(0.6)Sn(0.4)Sb2Te4 was confirmed by high-resolution electron microscopy and electron diffraction. In contrast, electron diffraction patterns of 9P-Ge(1.3)Sn(0.7)Sb2Te5 reveal a significant extent of stacking disorder as evidenced by diffuse streaks along the stacking direction. The Seebeck coefficient is unaffected by the Sn substitution but the thermal conductivity drops by a factor of 2 which results in a thermoelectric figure of merit ZT = ~0.25 at 450 °C for both Ge(0.6)Sn(0.4)Sb2Te4 and Ge(1.3)Sn(0.7)Sb2Te5, which is higher than ~0.20 for unsubstituted stable layered Ge-Sb-Te compounds.

Color-image quality assessment: from prediction to optimization.

While image-difference metrics show good prediction performance on visual data, they often yield artifact-contaminated results if used as objective functions for optimizing complex image-processing tasks. We investigate in this regard the recently proposed color-image-difference (CID) metric particularly developed for predicting gamut-mapping distortions. We present an algorithm for optimizing gamut mapping employing the CID metric as the objective function. Resulting images contain various visual artifacts, which are addressed by multiple modifications yielding the improved color-image-difference (iCID) metric. The iCID-based optimizations are free from artifacts and retain contrast, structure, and color of the original image to a great extent. Furthermore, the prediction performance on visual data is improved by the modifications.

Image-difference prediction: from grayscale to color.

Existing image-difference measures show excellent accuracy in predicting distortions, such as lossy compression, noise, and blur. Their performance on certain other distortions could be improved; one example of this is gamut mapping. This is partly because they either do not interpret chromatic information correctly or they ignore it entirely. We present an image-difference framework that comprises image normalization, feature extraction, and feature combination. Based on this framework, we create image-difference measures by selecting specific implementations for each of the steps. Particular emphasis is placed on using color information to improve the assessment of gamut-mapped images. Our best image-difference measure shows significantly higher prediction accuracy on a gamut-mapping dataset than all other evaluated measures.

Toward a unified color space for perception-based image processing.

Image processing methods that utilize characteristics of the human visual system require color spaces with certain properties to operate effectively. After analyzing different types of perception-based image processing problems, we present a list of properties that a unified color space should have. Due to contradictory perceptual phenomena and geometric issues, a color space cannot incorporate all these properties. We therefore identify the most important properties and focus on creating opponent color spaces without cross contamination between color attributes (i.e., lightness, chroma, and hue) and with maximum perceptual uniformity induced by color-difference formulas. Color lookup tables define simple transformations from an initial color space to the new spaces. We calculate such tables using multigrid optimization considering the Hung and Berns data of constant perceived hue and the CMC, CIE94, and CIEDE2000 color-difference formulas. The resulting color spaces exhibit low cross contamination between color attributes and are only slightly less perceptually uniform than spaces optimized exclusively for perceptual uniformity. We compare the CIEDE2000-based space with commonly used color spaces in two examples of perception-based image processing. In both cases, standard methods show improved results if the new space is used. All color-space transformations and examples are provided as MATLAB codes on our website.

From metastable to stable modifications-in situ Laue diffraction investigation of diffusion processes during the phase transitions of (GeTe)nSb2Te3 (6 < n < 15) crystals.

Temperature dependent phase transitions of compounds (GeTe)(n)Sb(2)Te(3) (n = 6, 12, 15) have been investigated by in situ microfocus Laue diffraction. Diffusion processes involving cation defect ordering at ~300 °C lead to different nanostructures which are correlated to changes of the thermoelectric characteristics.

Analyzing small suprathreshold differences of LCD-generated colors.

Small suprathreshold color differences around five CIE color centers were investigated on a typical liquid crystal display (LCD) with fluorescent backlight using the method of constant stimuli. The results were evaluated using probit analysis and compared with surface-color differences of the RIT-DuPont dataset. We focused especially on the relationship between T50 distances obtained from LCD and surface-color stimuli and on the influence of the display's narrowband primaries and its relatively low luminance level on interobserver uncertainty. The low luminance level of the LCD decreases the perceived color differences. However, considering the visual uncertainty of the experimental data, we could not reject the hypothesis that T50 distances from the RIT-DuPont and our experiment agree up to a constant scaling factor. In addition, we found significantly higher interobserver variability in the estimation of small color differences if the colors are viewed on an LCD. There are some indications that color-difference perception might be influenced by individual color-matching functions and, thus, by the spectral power distribution of the stimuli. We provide the experimental data, including all spectral stimuli shown to the observers, on our website.