Search

Disaster risk management of movable and immovable cultural heritage is a highly significant research topic. In this work, we present a pipeline for 3D digitisation, segmentation and annotation of large scale urban areas in order to produce data that can be exploited in disaster management simulators (e.g fire spreading, crowd movement, firefighting training, evacuation planning, etc.). We have selected the old town of Xanthi (Greece) as a challenging case study. We developed a custom multispectral camera to be carried by a commercial drone. Using the structure from motion / multiview stereo (SFM/MVS) approach, we produced a 3D model of the urban area covering 0.5 k m 2 that is followed by a multilayer texture map which carries information from visible and near-infrared regions of the electromagnetic spectrum. We developed a set of machine learning approaches based on logistic regression, support vector machines and artificial neural networks that allow 3D model segmentation by exploiting not only morphological and structural features but also the multispectral behaviour of different material surfaces. We objectively evaluate the performance of the proposed segmentation approaches on six significant material-based classes (cobbled-roads granite kilns, building walls, ceramic roof-tiles, low-vegetation, high-vegetation and metal surfaces) that are used in simulating fire propagation and crowd movement. The experiments revealed that the segmentation accuracy can be enhanced by taking into consideration surface material multispectral properties as well as morphological features. A Web-based multi-user annotation tool complements our proposed pipeline by enabling further 3D model segmentation, fine tuning and semantics annotation (e.g. usage-based building classification and evacuation priorities, escape paths and gathering points).

Texture mapping is a key step in large-scale 3D scene reconstruction, which can greatly enhance visual reality of the reconstructed scenes. However, existing techniques are unable to accomplish this task efficiently due to the high computational complexity of reconstructing large-scale real-world 3D scenes. In this work, we propose a new efficient convex optimization-based approach, i.e. the mesh-based continuous max-flow method, which can be easily implemented and accelerated upon a modern parallel computing platform, e.g. GPU. Particularly, an Markov Random Fields (MRF) based model, i.e. Potts model, is introduced to mathematically formulate the key specific view selection problem; we show that the challenging combinatorial optimization problem can be efficiently solved by resolving its convex relaxation, which recovers textures from images with the proposed duality-based continuous max-flow approach. In addition, visual effects of sharpness and deformation are utilized to define a criterion of evaluating texture quality effectively, and a large 3D triangular mesh is partitioned into structural components so as to reduce memory consumption of the proposed algorithm. The proposed mesh-based continuous max-flow approach for large-scale texture mapping demonstrates its outperformance over state-of-the-art methods, over large-scale public datasets, in both numerical efficiency and visual quality; meanwhile, our GPU-accelerated algorithm can yield 3D textured models with high quality from complex large-scale scenes in minutes.

During the annealing of Ni and Ni-Mo films, a {110} texture, usually considered to be both of high surface energy and high strain energy, was found to develop. Quantitative thermodynamic model calculations showed that the anomalous formation of this {110} texture originates fundamentally from the grain yielding anisotropy of Ni. Based on extensive grain yielding anisotropy model calculations, a “texture map” based on {111}, {100}, and {110} textures was constructed to predict the texture transition temperatures for different film thicknesses. A kinetic analysis of the texture evolution in the films is further presented, revealing that the texture evolution is controlled by the self-diffusion of atoms at grain boundaries. These findings pave the way for the achievement of unusual surface orientations through the quantitative texture design of thin films.

Nowadays, image processing and 3D shape analysis are an integral part of clinical practice and have the potentiality to support clinicians with advanced analysis and visualization techniques. The main contribution of the paper is the integration of these two approaches in order to increase the amount of information available and, thus, allow a more accurate analysis of each patient. Given a segmented anatomical district, we propose a novel mapping of volumetric data onto the segmented surface. The grey levels of the image voxels are mapped through a volume-surface correspondence map, which defines a grey level texture on the segmented surface. The resulting texture mapping is coherent to the local morphology of the segmented anatomical structure and provides an enhanced visual representation of the anatomical district. The integration of volume-based and surface-based information in a unique 3D representation also supports the identification and characterization of morphological landmarks and pathology evaluations.

Exemplar based texture synthesis is an important technique for image processing and computer graph in texture mapping. So far, great achievements have been made in this field. However, both traditional and modern methods based on deep learning are making errors in synthesizing patterned texture due to the failure for catching the regularity of texture. To obtain a better synthesized result, a new framework for regular texture synthesis is proposed in this paper. Besides, we use recurrent neural network (RNN) of long-shot term memory (LSTM) to produce a regular texture based on exemplar. Our method can generate at any size texture without errors, which is an improvement for texture synthesis with deep learning techniques.Compared with traditional method as well as deep learning method, our method is obviously better in synthesizing regular texture.

There are increasing demands nowadays on spatial soil information in order to support environmental related and land use management decisions. Physical soil properties, especially particle size distribution play important role in this context. A few of the requirements can be satisfied by the sand-, silt-, and clay content maps compiled according to global standards such as GlobalSoilMap (GSM) or Soil Grids. Soil texture classes (e. g. according to USDA classification) can be derived from these three fraction data, in this way texture map can be compiled based on the proper separate maps. Soil texture class as well as fraction information represent direct input of crop-, meteorological- and hydrological models. The model inputs frequently require maps representing topsoil features, which refer most commonly to 0–30 cm depth. This is covered by three consecutive depth intervals according to standard specifications: 0–5 cm, 5–15 cm, 15–30 cm. Becoming GSM and SoilGrids the most detailed freely available spatial soil data sources, the common model users (e. g. meteorologists, agronomists, or hydrologists) would produce input map from (the thickness-weighted mean of) these three layers. However, if the basic soil data and proper knowledge is obtainable, a soil texture map targeting directly the 0–30 cm layer could be independently compiled. In our work we compared Hungary's soil texture maps compiled using the same reference and auxiliary data and inference methods but for differing layer distribution. We produced the 0–30 cm clay, silt and sand map as well as the maps for the three standard layers (0–5 cm, 5–15 cm, 15–30 cm). Maps of sand-, silt-, and clay content were computed through composite regression kriging applying Additive Log-Ratio (alr) transformation. In addition to the Hungarian Soil Information and Monitoring System as reference soil data, digital elevation model and its derived components, soil physical property maps, remotely sensed images, land use-, geological-, as well as meteorological data were applied as auxiliary variables. We compared the directly compiled and the synthetized clay-, sand content, and texture class maps by different tools. In addition to pairwise comparison of basic statistical features (histograms, scatter plots), we examined the spatial distribution of the differences. We quantified the taxonomical distances of the textural classes, in order to investigate the differences of the map-pairs. We concluded that the directly computed and the synthetized maps show various differences. In the case of clay-, and sand content maps, the map-pairs have to be considered statistically different. On the other hand, the differences of the texture class maps are not significant. However, in all cases, the differences rather concern the extreme ranges and categories. Using of synthetized maps can intensify extremities by error propagation in models and scenarios. Based on our results, we suggest the usage of the directly composed maps.

Thermal image data are widely used to assess the insulation quality of buildings and to detect thermal leakages. In our approach, we merge terrestrial thermal image data and 3D point clouds to perform thermal texture mapping for building facades. Since geo-referencing data of a hand-held thermal camera is usually not available in such applications, registration between thermal images and a 3D point cloud (for instance generated from RGB image data by structure-from-motion techniques) is essential. In our approach, thermal image data registration is conducted in four steps: First, another point cloud is generated from the thermal image data. Next, a coarse registration between thermal point cloud and RGB point cloud is performed using the fast global registration (FGR) algorithm. The best corresponding thermal-RGB image pairs are acquired by picking up the lowest Euclidean distance between the exterior orientation parameters of thermal images and transformed exterior orientation parameters of RGB images. Subsequently, radiation-invariant feature transform (RIFT), normalized barycentric coordinate system (NBCS) and random sample consensus (RANSAC) are employed to extract reliable matching features on thermal-RGB image pairs. Afterwards, a fine registration is performed by mono-plotting of the RGB image, followed by image resection of the thermal image. Finally, in terms of texture mapping algorithms, in order to remove the blur effects caused by small misalignments for different candidate images, a global image pose refinement approach, which aims to minimize the temperature disagreements provided by different images for the same object points, is proposed. In addition, in order to ensure high geometric and radiant accuracy, camera calibrations are performed. Experiments showed that the proposed method could not only achieve high geometric registration accuracy, but also provide a good radiometric accuracy with RMSE lower than 1.5 °C.

This paper proposes a framework for reconstructing 3D human head models from a single image. Firstly, a preprocessing system for images is designed. The image is automatically segmented using convolutional neural network. The Gabor filter is used to extract the direction of the hair, and a optional manual interaction is used to modify the results. Secondly, the parameters of the FLAME face parametric model are solved by using landmark points as constraint. The high-frequency information of the face image is used to enhance the detail, and the texture map is complemented by the albedo parameterized model. Finally, a strip mesh hair database is constructed, and the hair in the image is reconstructed by using this database and the information extracted from the image.

Nowadays, the traditional measurement methods applied in ship sections/blocks are heavily relied on years of experiences and could hardly meet the shipbuilding requirements due to their complicated and time-consuming processes. Furthermore, along with the development of photogrammetry, the image-based three-dimensional (3D) reconstruction technology has emerged and become flourishing. In this paper, we propose a novel measurement method based on image-based 3D reconstruction technology to measure the construction deformation of ship blocks with high accuracy in a simple way. The scheme only needs photos to rebuild the 3D model of ship block. It includes the camera calibration, the image preprocessing, the feature points matching followed by the steps of dense point cloud generation, meshes partition and 3D reconstruction with texture mapping. In the step of feature points matching, the scale-invariant feature transform (SIFT) algorithm is used and optimized with the addition of the global vector to reduce the mismatch and enhance the reconstruction accuracy. The shipyard experiments with shooting photos of arbitrary steel parts proved that the proposed image-based 3D reconstruction technology is able to be applied to measure the ship blocks with satisfied accuracy and less time in shipbuilding industry or other industries with large-scale one-of-a-kind product as well.

With the development of mobile Internet and intelligent mobile devices, mobile iris recognition is a new advanced frontier for identification. This paper presents an iris recognition algorithm based on a multi-layer analogous convolutional structure and collaborative representation to solve the high intra-class difference caused by the visible lighting interference and the change of image acquisition sensors. The proposed method uses the multi-layer analogous convolution to reduce the dimension of iris texture information with lower computational complexity. Then, the feature of the obtained texture map is extracted and classified by collaborative representation method. The experimental analysis on MICHE-I demonstrates the adaptability of proposed method to changes in lighting condition and in acquisition sensors. The experimental results show that our algorithm provides better accuracy, AUC and EER compared to the iris recognition method such as Gabor, PCA, SRC, MACS + SRC and CRC.

We present an efficient approach for the parameterization of triangle soups. Our technique tackles the problem by first approximating the triangle soup by a watertight 2-manifold offset mesh proxy. By establishing point correspondences between the triangle soup and the shape proxy by tracing electrostatic field lines in time O(kn) instead of O(kn2), we can transfer properties computed on the surface, such as UV-coordinates even for large input meshes. The technique can process large triangle soups by replacing the computational intensive physical calculations by approximations developed initially for large-scale physical simulations. Thus, it is possible to decrease the time complexity of the initial charge distribution algorithm from O(n3) to O(n). We demonstrate our method on a multitude of challenging triangle soups and focus on the transfer of UV-coordinates during our experiments. An intensive run-time analysis and a comparison of our results to state of the art techniques in standard modeling tools concludes our study.

Globally unified normal orientations for unorganized point set are essential for reconstruction, rendering, texture mapping etc. Although there is a large body of literature on dealing with various difficult cases to get consistent normal orientation, effective approaches are not available yet to handle a point cloud with thin surfaces, to our best knowledge. The difficulty lies in that thin structure doesn’t have a strong coherence between locations of surface points and their normal orientations. In this paper, we propose a novel approach for computing globally optimal normal orientations to facilitate high-quality surface reconstruction from a point cloud with thin structures. The key idea is to generate an uniform particle set constrained on an offset surface that is topologically identical to the real target surface and further use the particle set to compute and refine the normal vector at each primitive point. Our strategy for inferring the globally consistent orientation is much different from the conventional paradigm that computes unoriented normal vectors and then re-orients them coherently. Experimental results on natural and man-made models with noise and thin parts show that our algorithm outperforms existing approaches on normal consistence and is able to produce a desirable global optimal normal orientations with which a structurally sound thin shape can be finally reconstructed by the well known screened Poisson reconstruction approach.

Computing geodesic distance on a mesh surface S efficiently and accurately is a central task in numerous computer graphics applications. In order to deal with high-resolution mesh surfaces, a lightweight preprocessing is a proper choice to make a balance between query accuracy and speed. In the preprocessing stage, we build a proximity graph G with regard to a set of sample points and keep the exact geodesic distance between any pair of nearby sample points. In the query stage, given two query points s and t , we augment the proximity graph G by adding s and t on-the-fly, and then use the shortest path between s and t on the augmented proximity graph to approximate the exact geodesic path between s and t . We establish an empirical relationship between the number of samples and expected accuracy (measured in relative error), which facilitates fast and accurate query of geodesic distance with a lightweight processing cost. We exhibit the uses of the new approach in two applications—real-time computation of discrete exponential map for texture mapping and interactive design of spline curves on surfaces.

Parameterization of triangulated surface meshes is a crucial problem in computer graphics, computer aided geometric design and digital geometric processing. This paper addresses the problem of planar parameterization, i.e., mapping a given triangulated surface onto a planar domain. We construct an optimized algorithm for parameterization of genus-zero meshes and aim to minimize the distortion of the parameterization. An energy functional is proposed in the paper, that quantities angle and area distortions simultaneously, while the relative importance between angle and area preservation can be controlled by the user through a parameter. The method is based on an iterative procedure that incrementally flattens mesh by growing region to obtain a parameterization result with free boundary. The result is then converted to a parameterization with regular boundary by conformal mapping. Application of the method to texture mapping is presented. Experiments show that the proposed method can obtain better results than some common parameterization methods.

Accurately recording information is the single most important stage of an archaeological project. The biggest technological improvement to documentation techniques in the last 15 years has been the spread of various 3D digitization technologies, such as computer vision photogrammetry and 3D laser scanning. These technologies have allowed archaeologists to quickly and accurately capture and reconstruct the geometry and colors of the subjects being studied. Each of the various methods for 3D digitization of cultural heritage materials has advantages and disadvantages, such as processing speed, cost of equipment, and the accuracy of the captured data. Laser scanned data is among the most accurate geometrical data available in modern scanning techniques, but it lacks in its ability to accurately capture textures and diagnostic coloration information. Photogrammetric data produces highly detailed photographic textures on models, but the geometric data tends to be of a lower definition than the laser scanned data. In this paper, the authors discuss a new methodology that combines the advantages of computer vision photogrammetry with those of laser scanning by applying the photographic textures produced with photogrammetry to the geometric data obtained from laser scanning. This method allows archaeologists to achieve the best possible fidelity 3D models for interpretation and study.

This paper proposes a novel local/global spherical parameterization for the genus-zero triangular mesh, which naturally extends the planar approach to the spherical case. In our method, we derive two fitting matrices (conformal and isometric) in 3D space. By optimizing the so-called spring energy, the spherical results are achieved by solving a nonlinear system with spherical constraints. Intuitively, it represents the stitching together of the 1-ring patches to form a unit sphere. Moreover, the derivation of the 3D fitting matrices can also be applied to planar triangles directly, so that we can obtain a class of novel planar approaches (conformal, isometric, authalic) to the problem of flattening triangular meshes. In order to enhance robustness of the proposed spherical method, a stretch operator is introduced for dealing with high-curvature models. Numerical results demonstrate that our method is simple, efficient and convergent, and it outperforms several state-of-the-art methods in terms of trading-off the distortions of angle, area and stretch. Furthermore, it achieves better visualization in texture mapping.

Recent years witnessed that deep learning based methods have achieved significant progresses in recovering 3D face shape from single image. However, reconstructing realistic 3D facial texture from single image is still a challenging task due to the unavailability of large-scale training datasets and the low expression ability of previous statistical texture models (e.g. 3DMM). In this paper, we introduce a novel deep architecture trained by self-supervision with multi-view setup, to reconstruct 3D facial texture. Specifically, we first obtain incomplete UV texture map from input facial image, and then introduce a Texture Completion Network (TC-Net) to inpaint missing areas. To train TC-Net, firstly, we collect 50,000 triplets of facial images from in-the-wild videos, each triplet consists of a nearly frontal, a left-side, and a right-side facial images. With this dataset, we propose a novel multi-view consistency loss that ensures consistent photometric, face identity, 3DMM identity, and UV texture among multi-view facial images. This loss allows to optimize TC-Net in a self-supervision way without using ground-truth texture map as supervision. In addition, multi-view input images are only required in training to provide self-supervision, and our method only needs single input image in inference. Extensive experiments show that our method achieves state-of-the-art performance in both qualitative and quantitative comparisons.

Starches are commonly used in sauces, dressings, dairy products, bakery items, confections, and snacks due to their wide range of functional properties. However, using starches to deliver palatable textures is often challenging because multiple texture and rheological attributes have to be balanced simultaneously. Textural properties of even simple starch-thickened foods can vary in surprising ways with changes in starch swelling volume and concentration. Thus, we have modeled the swelling behavior of starch thickeners to develop rheological maps of starch texture as a predictive tool for formulation of starch-thickened foods. Predictive equations for starch suspension viscosity, elastic modulus, phase angle, critical strain and shear thinning index were developed as a function of starch swelling volume and concentration, and these equations showed good fit with experimental data (R2 ≥ 0.89). The developed starch texture maps overlay cutoff lines of various texture behaviors over plots of either viscosity or elastic modulus as a function of starch swelling volume and concentration, indicating zones of various rheologically-influenced texture properties such as shear-thinning and significant elastic behavior. The starch texture maps provide a powerful predictive and visual tool that enables targeted, efficient development of texture solutions for starch-thickened foods.

In this paper, we propose a phase-field model to partition a curved surface into path-connected segments with minimal boundary length. Phase-fields offer a powerful tool to represent diffuse interfaces with controlled width and to optimize them in a variational framework. We demonstrate how the multiplicative combination of phase-field functions can be used to effectively compute a hierarchical partition of unity. This induces an associated hierarchy of atlases, whose charts naturally overlap and thus are well-suited for applications such as texture mapping. Furthermore, we obtain distortion minimizing segmentations via a PDE-constraint optimization approach where the phase-field model allows direct use of Lagrangian calculus. Following Sharp and Crane (2018) , the Yamabe equation, which allows computing the distortion induced by segment flattening, is considered as the constraint. This way, we obtain end-to-end diffuse formulations of variational problems in surface segmentation that are straightforward to treat computationally. Various examples illustrate the flexibility and robustness of this approach.

High-speed three-dimensional (3D) shape measurement techniques based on fringe projection profilometry (FPP) have undergone huge advances over the past two decades. However, accurate 3D displacement mapping and deformation analysis of dynamic scenes using FPP remains an unsolved problem. Because fringe patterns are projected rather than attached on the tested surfaces, the full-field point-to-point correspondence cannot be accurately established between any two 3D shape results. To deal with this challenge, a DIC-assisted FPP for high-speed 3D shape, displacement and deformation measurement of textured surfaces is proposed. Firstly, a high-speed 3D shape measurement system is adopted using our recently proposed robust and efficient Gray-coded coding strategy, which can accurately reconstruct full-field shape of discontinuous surfaces with rich texture information. Then, the modulation-based method is proposed to retrieve high-quality texture map from three phase-shifting fringe patterns, which can eliminate the adverse influence of the nonuniform and time-varying ambient light. By matching the retrieved surface texture images at difference states using DIC, accurate point tracking between the measured 3D shape data can be fulfilled, leading to precise 3D displacement and deformation measurement. Experiments have verified that the proposed method can achieve 3D shape, displacement and deformation measurement of dynamic scenes at a frame rate of 542 fps without any extra expense of hardware or calibration for the FPP system. The presented method is reliable and promising for further 3D displacement mapping and deformation analysis of dynamic scenes using FPP.

Stenosis refers to the thinning of the inner surface (lumen) of vascular structures. Detecting stenoses and correctly estimating their degree is crucial in clinical settings for proper treatment planning. Such a planning involves a visual assessment, which in case of vascular structures is frequently based on 3D visual representations of the vessels. However, since vessel segmentation is affected by various sources of errors and noise in the imaging and image processing pipeline, it is crucial to capture and visually convey the uncertainty in a 3D visual representation. Moreover, it is crucial to quantify how much this uncertainty affects the calculated stenotic degree, since different severities lead to different treatments. We propose a novel approach for visualizing the shape deviation of different probability levels in vascular data, where the probability levels are computed from a probabilistic segmentation approach. Our non-obstructive visual encoding is based on rendering a single opaque surface representing a probability level of the cumulative distribution function around the vessels’ centerline. The surface rendering is enhanced with cumulative information about other levels. To do so, we traverse the probability space by applying an iterative projection method both inwards and outward until we reach surface variability within a given margin. We capture the shape variability between the different probability levels using the lengths of the projection lines, the change in angular directions, and the distortion of a parametrization. They are visually encoded using color and texture mapping. Furthermore, we allow for an interactive selection of a region of interest that automatically calculates the stenotic degree and how much the uncertainty affects the most likely result. We analyze our approach in comparison to state-of-the-art methods with medical experts in a study using both real magnetic resonance (MR) and computed tomography (CT) angiography data of vertebral arteries with stenoses as well as on MR angiography data with synthetically added stenoses and stenotic uncertainties. We evaluate how well our approach can guide medical experts in their assessment of the uncertainty in vertebral stenoses.

In this paper, we propose an unsupervised salient object segmentation approach using saliency and object features. In the proposed method, we utilize occlusion boundaries to construct a region-prior map which is then enhanced using object properties. To reject the non-salient regions, a region rejection strategy is employed based on the amount of detail (saliency information) and density of KAZE keypoints contained in them. Using the region rejection scheme, we obtain a threshold for binarizing the saliency map. The binarized saliency map is used to form a salient superpixel cluster. Finally, an iterative grabcut segmentation is applied with salient texture keypoints (SIFT keypoints on the Gabor convolved texture map) supplemented with salient KAZE keypoints (keypoints inside saliency cluster) as the foreground seeds and the binarized saliency map (obtained using the region rejection strategy) as a probably foreground region. We perform experiments on several datasets and show that the proposed segmentation framework outperforms the state of the art unsupervised salient object segmentation approaches on various performance metrics. Display Omitted Effective object segmentation using saliency and KAZE features is proposed.Region rejection strategy utilizing saliency and density of KAZE keypoints.KAZE keypoints are most suited for characterization of boundaryness.Objectness level information is enhanced with the help of salient keypoints.Outperform state of the art unsupervised salient object segmentation techniques.

This paper presents an image-based object reconstruction with a low memory footprint using run-length representation. While conventional volume-based approaches, which utilize voxels as primitives, are intuitive and easy to manipulate 3D data, they require a large amount of memory and computation during the reconstruction process. To overcome these burdens, this paper uses 3D runs to represent a 3D object and reconstructs each 3D run from multi-view silhouettes with a small amount of memory. The proposed geometry reconstruction is also computationally inexpensive, as it processes multiple voxels simultaneously. And for the compatibility with the conventional data formats, generation of polygonal 3D meshes from the reconstructed 3D runs is proposed as well. Lastly, texture mapping is proposed to additionally reduce the amount of memory for object reconstruction. The proposed reconstruction scheme has been simulated using various types of multi-view datasets. The results show that the proposed method performs object reconstruction with a smaller amount of memory and computation than voxel-based approaches. An image-based object reconstruction using run-length representation is proposed.A fast geometry reconstruction by rectifying images is proposed.A 3D mesh generation algorithm from the reconstructed 3D runs is proposed.View dependent texture mapping algorithm using a color palette is proposed.

Orthogonal projections of objects play an important role in the process of making archaeological illustrations. We present a method to generate the detailed orthogonal projection of a 3D model by fusing normal and texture information in gradient domain. We first render the model into a texture image from a perpendicular view. A normal map is then obtained from the same view with pseudocolors converted from vertex normals. Finally, we make a non-photorealistic projection image that combines the texture image and the normal map by solving the 2D screened Poisson equation. The non-photorealistic projection is both geometry-aware and texture-aware and enhances the subtle details that are hard to see in the texture image or the normal map alone. It is more convenient for archaeologists to make line drawings by using tools such as Adobe Illustrator ® to trace over the fused images.

To construct an intelligent space with a distributed camera sensor network, pre-calibration of all cameras (i.e., determining the absolute poses of each camera) is an essential task that is extremely tedious. This paper considers the automatic calibration method for camera sensor networks based on 3D texture map information of a given environment. In other words, this paper solves a global localization problem for the poses of the camera sensor networks given the 3D texture map information. To manage the complete calibration problem, we propose a novel image descriptor based on quantized line parameters in the Hough space (QLH) to perform a particle filter-based matching process between line features extracted from both a distributed camera image and the 3D texture map information. We evaluate the proposed method in a simulation environment with a virtual camera network and in a real environment with a wireless camera sensor network. The results demonstrate that the proposed system can calibrate complete external camera parameters successfully.

For better removing textures with larger gradients while preserving structural edges with smaller gradients, we propose an image smoothing method based on histogram equalized content-aware patches and direction-constrained sparse gradients. In order to better smooth the boundary concentration regions while maintaining the continuity of boundary pixels, a content-aware patching technology with boundary constraints is proposed. The irregular patches are represented by the smallest rectangular bounding boxes to reduce the computational complexity. Based on edge information, patches are divided into edge-patches and non-edge-patches. Histogram equalization is used to improve the edge gradient of patches with structural edge concentration, while the image decomposition is used to reduce the gradient of texture details. Taking the edge information as the smoothing factor, each patch is smoothed via direction-constrained sparse gradients. The entire image needs to be further smoothed to remove residual texture details. Experimental results show that new method has better visual effects in retaining structural edges and removing texture details, and has many applications, including edge detection, image abstraction, detail enhancement, and texture mapping.

Folk's textual classification scheme which is widely used for sediment study operates with the proportions of gravel, sand, silt and clay fractions conventionally. However, dealing with data from different sources usually needs to face missing values that may make the classification difficult. To solve this problem and discover other methods of analyzing the scheme, with samples of offshore seabed sediment, a two-stage model was established to predict a sample's class using the XGBoost algorithm as well as the grain size parameters as input features. The final model was evaluated with quantitative performance measures of recall, precision and F1 score, and by comparing sediment texture maps using the predicted and the actual data. The results show that the model performs well on extraction of sediment samples without gravel fraction, and prediction of classes that have independent characteristics of grain size parameters or samples not near the boundaries of classes in the ternary diagram. The predicted sediment texture is close to the actual and could be reliable due to errors with little impact on further applications. It is demonstrated that the model could be an auxiliary or alternative approach to offshore sediment texture mapping, as well as supplementary to the analysis of sedimentary environment.

In recent years GPUs have become one of the most important components in mobile application processors (APs). Thus, performance measurement and analysis of mobile GPUs are crucial to mobile AP manufacturers, device manufacturers, graphics application programmers, and end users. However, it is hard to analyze mobile GPUs in depth via existing high-level (with frames per second) or low-level benchmarks (with a fill rate, ALU performance, etc.). To bridge the gap between the benchmarks, we present a novel Android benchmark set for low-power GPUs, called L-Bench. This benchmark set consists of mid-level micro-benchmarks implemented on OpenGL ES 3.1, which are carefully chosen for different workloads. By analyzing the results, this benchmark suite provides not only frames per second of each benchmark but also performance of each GPU subsystem (geometry units, ALUs, texture mapping units, raster operations pipelines, caches/memory units, and tessellators) and overall GPU performance. For experiments, we perform our benchmark suite on five representative mobile devices that have different mobile GPUs, after that, we describe comprehensive analysis of each GPU architecture.

Autonomous vehicles have a large number of on-board sensors, not only for providing coverage all around the vehicle, but also to ensure multi-modality in the observation of the scene. Because of this, it is not trivial to come up with a single, unique representation that feeds from the data given by all these sensors. We propose an algorithm which is capable of mapping texture collected from vision based sensors onto a geometric description of the scenario constructed from data provided by 3D sensors. The algorithm uses a constrained Delaunay triangulation to produce a mesh which is updated using a specially devised sequence of operations. These enforce a partial configuration of the mesh that avoids bad quality textures and ensures that there are no gaps in the texture. Results show that this algorithm is capable of producing fine quality textures. Texture mapping of Geometric Scene Representations.Incremental mapping of texture from continuous throughput of sensor data.Texture enhancement from multiple images.

Seebeck coefficient values measured for textured polycrystalline Bi films have been compared with numerically computed ones. The proposed computation procedure requires crystallographic texture mapping and values of single-crystalline transport coefficients as input data. The comparison of computed and measured Seebeck coefficients allowed determining the factors affecting films' transport properties except of crystalline texture. Two particular cases are considered: when the film properties are affected only by texture and when they depend on texture and grain size simultaneously.

In order to support an effective planning of solar panel constructions or passive energy measures, information of potential solar income has to consider multiple levels of scale and detail. It has to consider the regional and the detailed household scale. Recent developments in the field of digital city models and 3D GIS allow combining the advantages of both domains. We present a multi-scale 3D GIS approach for the assessment and dissemination of solar income of a digital city model. For the consideration of long and close range shadows, we combine polygonal geometries, 3D-voxel grids and 2.5D Digital Elevation Models of different resolutions in a multi-scale modelling scene. In order to allow a realistic assessment of the heterogeneous 3D shadowing patterns in urban environments, we present a texture mapping approach for homogeneous surface sampling and efficient data storage. The approach integrates the regional topographic context with detailed 3D assessments on a household scale. In order to show the capabilities of the presented approach, a performance test using different scene compilations was conducted. Comparing the performance of different geometrical representations of a city model tile in terms of computational effort and accuracy of the results, the relevance of high spatial resolutions and the consideration of full 3-dimensionality of the presented modelling approach is proven. The results are compatible to current 3D standards such as CityGML and COLLADA facilitating the dissemination of results to non-expert addressees.

In this paper, we explore texture mapping as a unified representation for enabling realistic multimodal interaction with finely detailed surfaces. We show how both normal maps and relief maps can be adopted as unified representations to handle collisions with textured rigid body objects, synthesize complex sound effects from long lasting collisions and perform rendering of haptic textures. The resulting multimodal display system allows a user to see, hear, and feel complex interactions with textured surfaces. By using texture representations for seamlessly integrated multimodal interaction instead of complex triangular meshes otherwise required, this work is able to achieve up to 25 times performance speedup and reduce up to six orders of magnitude in memory storage. We further validate the results through user studies to demonstrate the effectiveness of texture representations for integrated multimodal display. Graphical abstractDisplay Omitted HighlightsWe present a system displaying fine details of visuals, haptics, and audio.We present a new method for rigid-body physics on normal-mapped surfaces.Multimodal interaction with texture maps reduces sensory conflict.Users prefer multimodal relief map interaction over normal maps.

High-speed real-time three-dimensional (3D) shape measurement with color texture has a large potential in various fields. The majority of the existing methods retrieve the 3D shape with color texture by employing an RGB camera or additional infrared camera; however, the RGB camera is more expensive and has lower sensitivity to light, and employing additional infrared camera requires complex alignment. Given this, a novel method is proposed to simultaneously retrieve the 3D shape and color texture. Specifically, compared with the existing methods, the proposed method only uses one monochromatic camera; as a result, the cameras alignment is not required, simplifying the 3D measurement system and reducing its cost. Three fringe patterns with different frequencies and three corresponding uniform gray-level patterns are set to different colors of red, green, and blue. Then Fourier transform profilometry (FTP) is employed to retrieve the 3D shape from the captured six images. While three uniform gray-level patterns can be used to eliminate the effects of the zero frequency term, they can be also used to generate the color texture of objects meanwhile. The proposed method only requires the capture of six monochromes, and the experimental results demonstrate the success of the proposed method.

The lunar global texture maps of roughness and entropy are derived at kilometer scales from Digital Elevation Models (DEMs) data obtained by Lunar Orbiter Laser Altimeter (LOLA) aboard on Lunar Reconnaissance Orbiter (LRO) spacecraft. We use statistical moments of a gray-level histogram of elevations in a neighborhood to compute the roughness and entropy value. Our texture descriptors measurements are shown in global maps at multi-sized square neighborhoods, whose length of side is 3, 5, 10, 20, 40 and 80 pixels, respectively. We found that large-scale topographical changes can only be displayed in maps with longer side of neighborhood, but the small scale global texture maps are more disorderly and unsystematic because of more complicated textures' details. Then, the frequency curves of texture maps are made out, whose shapes and distributions are changing as the spatial scales increases. Entropy frequency curve with minimum 3-pixel scale has large fluctuations and six peaks. According to this entropy curve we can classify lunar surface into maria, highlands, different parts of craters preliminarily. The most obvious textures in the middle-scale roughness and entropy maps are the two typical morphological units, smooth maria and rough highlands. For the impact crater, its roughness and entropy value are characterized by a multiple-ring structure obviously, and its different parts have different texture results. In the last, we made a 2D scatter plot between the two texture results of typical lunar maria and highlands. There are two clusters with largest dot density which are corresponded to the lunar highlands and maria separately. In the lunar mare regions (cluster A), there is a high correlation between roughness and entropy, but in the highlands (Cluster B), the entropy shows little change. This could be subjected to different geological processes of maria and highlands forming different landforms.

Sculptures, one of the most important forms of cultural heritage, exist in numerous temples, towers and grottoes. The current generation of high precision 3D digital sculpture models is a primary reality-based modeling tool for digital documentation, archaeological research, virtual restoration, and digital exhibition. However, high precision texture mapping remains a challenge in high fidelity digital reconstruction of complex sculptures, especially in the areas of rich geometrical details. This paper proposes a method to texture mapping one image onto the sculpture model to deal with the mapping problem problems caused by the image distortion and the model deformation. Firstly, we use a rigid pinhole camera model to project the image onto the geometric model to acquire primary texture mapping, where direct linear transformation is employed to estimate the elements of exterior orientation of the texture image. Then, we propose a non-rigid transformation model, called weighted thin-plate spline (W-TPS), discuss the theoretical derivation and the establishment of the W-TPS function, and apply this model to adjust the primary texture mapping results. The proposed method can precisely project the image onto the geometrical model both globally and locally. We conduct three experiments to verify the efficiency and feasibility of the proposed method using the sculptures in Dunhuang Mogao Grottoes in western China, which was inscribed on World Heritage List by UNESCO in 1987.

This paper investigates the effects of adding texture to images with poorly-textured regions on optical flow performance, namely the accuracy of foreground boundary detection and computation time. Despite significant improvements in optical flow computations, poor texture still remains a challenge to even the most accurate methods. Accordingly, we explored the effects of simple modification of images, rather than the algorithms. To localize and add texture to poorly-textured regions in the background, which induce the propagation of foreground optical flow, we first perform a texture segmentation using Laws' masks and generate a texture map. Next, using a binary frame difference, we constrain the poorly-textured regions to those with negligible motion. Finally, we calculate the optical flow for the modified images with added texture using the best optical flow methods available. It is shown that if the threshold used for binarizing the frame difference is in a specific range determined empirically, variations in the final foreground detection will be insignificant. Employing the texture addition in conjunction with leading optical flow methods on multiple real and animation sequences with different texture distributions revealed considerable advantages, including improvement in the accuracy of foreground boundary preservation, prevention of object merging, and reduction in the computation time. The F-measure and the Boundary Displacement Error metrics were used to evaluate the similarity between detected and ground-truth foreground masks. Furthermore, preventing foreground optical flow propagation and reduction in the computation time are discussed using analysis of optical flow convergence.

Depth estimation is a classical problem in computer vision and after decades of research many methods have been developed for 3D perception like magnetic tracking, mechanical tracking, acoustic tracking, inertial tracking, optical tracking using markers and beacons. The vision system allows the 3D perception of the scene and the process involves: (1) camera calibration, (2) image correction, (3) feature extraction and stereo correspondence, (4) disparity estimation and reconstruction, and finally, (5) surface triangulation and texture mapping. The work presented in this paper is the implementation of a stereo vision system integrated in humanoid robot. The low cost of the vision system is one of the aims to avoid expensive investment in hardware when used in robotics for 3D perception. In our proposed solution, cameras are highly utilized as in our opinion they are easy to handle, cheap and very compatible when compared to the hardware used in other techniques. The software for the automated recognition of features and detection of the correspondence points has been programmed using the image processing library OpenCV (Open Source Computer Vision) and OpenGL (Open Graphic Library) is used to display the 3D models obtained from the reconstruction. Experimental results of the reconstruction and models of different scenes are shown. The results obtained from the program are evaluated comparing the size of the objects reconstructed with that calculated by the program. Implementation of a stereo vision system integrated in humanoid robot is proposed.Low cost robotics vision system for 3D perception avoids expensive hardware cost.Cameras are highly utilized as they are easy to handle, cheap and very compatible.

We propose a color adjustment technique to eliminate the visible seams in image-based texture map of a 3D object. The process is carried out in three steps. First, texture coordinates are locally displaced to minimize the misalignment of adjacent texture patches. Second, color discontinuities between different texture patches at each corner of the mesh faces are resolved. We minimize a global energy function over the mesh to ensure continuous color transitions and fit the color gradient at each corner of the mesh faces. Finally, the color adjustment at the corners is propagated over the texture patch for each face by solving a Poisson equation with mixed boundary conditions. By means of the proposed processing techniques, the visibility of seams is minimized while fine details are preserved in image-based texture maps. This can be used as a last refinement stage in image-based 3D reconstruction pipelines. The proposed color adjustment algorithm is tested on a variety of real-world datasets and compares very favorably with known methods. Display Omitted

Treatment using implants is frequently employed in prosthetic dentistry. In this method, determining the bone density of the upper and lower jaws is important. Generally, a dentist can recognize the condition of the alveolar bone to be manipulated using a cone-beam computed tomography (CBCT) image. However, communicating the data to the patient is a challenge because it is difficult for the nonprofessional person to interpret the image, which contains a distribution of pixels with similar density. We present an intuitive texture mapping method of the alveolar bone area for application in implant treatment. Our method aims to help patients better understand the treatment process by using a textured image that includes several different texture patterns that reflect the density of the alveolar bone area. We segment the area in accordance with the density of corresponding parts in the alveolar bone and the gingiva. By simplifying the boundary of each segmented region, the distribution of pixels with similar density on the alveolar bone area can be easily recognized. Next, the texture patterns for several segmented regions are mapped onto the alveolar bone area using the graph-cut algorithm, which is used for smooth texture mapping at the boundary of the segmented region. The result is an applied texture on the alveolar bone area that corresponds to the bone structure. Our method is helpful for facilitating communication and understanding of treatment using dental implants. HighlightsWe propose intuitive texture mapping on the alveolar bone area for implant treatment.After segment alveolar bone, texture patterns are mapped using graph-cut algorithm.We make result applying a texture image corresponding to the bone structure.It is helpful for facilitating communication and understanding of implant treatment.

Alignment is a key step of the point cloud processing. We propose a flexible, easy-implemented, low-cost alignment method for multi-view point clouds of small-size objects (≦50 cm). The method is based on the idea of rotating measurement, firstly more than three pieces of point clouds of cylinder calibration pattern are used for calibrating the center axis of a single-axis turntable in the scanner’s coordinate system, so the rigid relationship of the turntable and scanner can be obtained. Then the rotation matrix can be calculated using the known angle parameter of the turntable, and finally an arbitrary view of point cloud can be easily transformed into a unified coordinate system by the rotation transformation. Our method does not require overlapping feature or co-markers constraint between point clouds, and only the rigid relationship between the scanner and the turntable and rotate angle need to be known. More importantly, the calibration of the turntable is flexible and low-cost: (1) the calibration pattern can be a common cylinder objects (such as cup, paper tube, caddy, water pipe, etc.), (2) the turntable and the alignment method can be integrated with variety 3-D scanners. The result of experiment has demonstrated the efficiency and automatic characteristic of the alignment method compared with other two convention methods.

Affine invariant feature extraction has been one of the key issues for object recognition, especially for the images captured under the variable environments. Considering that multiscale autoconvolution feature (MSA), which has the prominent comprehensive performance, is very sensitive to illumination change, a novel algorithm of extracting affine invariant feature is proposed based on the MSA transform combining with texture structure analysis. Firstly, a new MSA feature is extracted from texture structure map of the image which is computed based on local binary pattern theory. And then the original image based MSA and the texture map based MSA are combined to a new feature using generalized canonical correlation analysis, called TFMSA. This new feature represents much more image information than the traditional one and is performed in various object recognition tasks. The experimental results indicate that the new TFMSA not only conquers the defect of the traditional MSA, but also has good adaptability for a certain range of viewing angles, partial occlusion, uniform and non-uniform illumination changes. The recognition accuracy of the new feature is superior to MSA and other improved methods.

Interactively mapping texture into a curved surface has wide applications. Apart from mapping texture to desired positions with low distortions, which has been considered by existing works, few work considered the intuitive need to distribute distortions according to the texture contents. In this work, we present a texture mapping method guided by importance map to preserve the shape of the prominent content. We formulate it as an importance-value-weighted parameterization. The mapping distortions are measured by LSCM+ energy, which is capable of decreasing the appearance of shrunken and fold-over triangles as well as shape preservation; and the weights are efficiently calculated by transforming the area integral into a line integral. To solve the parameterization, we employ the ‘L–M’ (Levenberg–Marquardt) method and alternately update the weights and the coordinates since they are dependent on each other. Finally, we show some examples to demonstrate the content-aware effect by comparison to traditional parameterization.

This paper presents an automatic extraction method for discontinuity trace mapping from 3D point cloud of natural rocky slopes. Unlike the methods based only on geometrical information, the proposed method also takes textural information into account. By using the texture mapping method, a texture map and a normal map are generated from the point cloud, which contain textural and geometrical information respectively. Then, the rolling guidance filter and a multi-scale edge chain detector are applied to the image to extract discontinuity traces. Finally, edge chains in the image were projected back to the 3D coordinate system through the previously established texture map. The proposed method is applied to three case studies, each representing a different situation. The results show that the proposed discontinuity trace extraction method is automatic, robust, extensible and performs better than geometry-based methods for smooth rock surfaces along which the curvature changes are not apparent. The proposed method could be used as a supplement to traditional contact discontinuity trace mapping methods.

Defocus blur in images deteriorates the quality of image-based textures for 3D archaeological models. We propose a method to estimate the defocus blur in calibrated multi-view images based on photographic optics. Our method estimates the defocus blur maps without any modification to cameras or using any statistical model of images. After the focus distance of each image is estimated from the reconstructed 3D models with physical scale, the near distance dn and the far distance df of acceptable sharpness from the view point are calculated. We set the colour of the model vertex to white if the vertex’s distance from the view point is within the interval between dn and df, otherwise black. Finally we render the 3D model into a defocus map using the intrinsic and extrinsic parameters of the camera. The obtained defocus maps can be utilized as masks to generate an all-focused texture map for the reconstructed 3D archaeological model. Experimental results on real data demonstrate the effectiveness of the method.

The demand for 3D city-scale models has been significantly increased due to the proliferation of urban planning, city navigation, and virtual reality applications. We present an approach to automatically reconstruct buildings densely spanning a large urban area. Our method takes as input calibrated aerial images and available GIS meta-data. Our computational pipeline computes a per-building 2.5D volumetric reconstruction by exploiting photo-consistency where it is highly sampled amongst the aerial images. Our building surface graph cut method overcomes errors of occlusion, geometry, and calibration in order to stitch together aerial images and yield a visually coherent texture-mapped result. Our comparisons show similar quality to the manually modeled buildings of Google Earth, and show improvements over naive texture mapping and over space-carving methods. We have tested our algorithms with a 12sqkm area of Boston, MA (USA), using 4667 images (i.e., 280GB of raw image data) and producing 1785 buildings.

In this paper, the accurate method for texture reconstruction with non-desirable moving objects into dynamic scenes is proposed. This task is concerned to editor off-line functions, and the main criteria are the accuracy and visibility of the reconstructed results. The method is based on a spatio-temporal analysis and includes two stages. The first stage uses a feature points tracking to locate the rigid objects accurately under the assumption of their affine motion model. The second stage involves the accurate reconstruction of video sequence based on texture maps of smoothness, structural properties, and isotropy. These parameters are estimated by three separate neural networks of a back propagation. The background reconstruction is realized by a tile method using a single texton, a line, or a field of textons. The proposed technique was tested into reconstructed regions with a frame area up to 8-20%. The experimental results demonstrate more accurate inpainting owing to the improved motion estimations and the modified texture parameters.

3D model segmentation avails to skeleton extraction, shape partial matching, shape correspondence, texture mapping, shape deformation, and shape annotation. Many excellent solutions have been proposed in the last decade. How to efficiently evaluate these methods and impartially compare their performances are important issues. Since the Princeton segmentation benchmark has been proposed, their four representative metrics have been extensively adopted to evaluate segmentation algorithms. However, comparison to only a fixed ground-truth is problematic because objects have many semantic segmentations, hence we propose two novel metrics to support comparison with multiple ground-truth segmentations, which are named Similarity Hamming Distance (SHD) and Adaptive Entropy Increment (AEI). SHD is based on partial similarity correspondences between automatic segmentation and ground-truth segmentations, and AEI measures entropy change when an automatic segmentation is added to a set of different ground-truth segmentations. A group of experiments demonstrates that the metrics are able to provide relatively higher discriminative power and stability when evaluating different hierarchical segmentations, and also provide an effective evaluation more consistent with human perception.

The automatic process for texturing 3D urban environments realistically is a challenging topic in Computer Graphics, especially when the model is generated from 2D GIS data from real sites. In order to achieve realistic scenes, buildings, urban furniture and street roadways should be modeled automatically also considering the slope of steep streets. A texturing process by hand which takes into account all the urban elements of a city is an impracticable task, especially if quality results are required. In this paper we propose an automatic method for building texturization in entire urban models using a small set of images. The different features of lower and upper parts of buildings are considered by implementing two different genetic algorithms. Our method can be used in cities with both horizontal and steep streets because the street slope is processed by the algorithm.

In this work, we present the concept of "Hamiltonian cycle art" that is based on the fact that any mesh surface can be converted to a single closed 3D curve. These curves are constructed by connecting the centers of every two neighboring triangles in the Hamiltonian triangle strips. We call these curves surface covering since they follow the shape of the mesh surface by meandering over it like a river. We show that these curves can be used to create wire sculptures and duotone (two-color painted) surfaces.To obtain surface covering wire sculptures we have developed two methods to construct corresponding 3D wires from surface covering curves. The first method constructs equal diameter wires. The second method creates wires with varying diameter and can produce wires that densely cover the mesh surface.For duotone surfaces, we have developed a method to obtain surface covering curves that can divide any given mesh surface into two regions that can be painted in two different colors. These curves serve as a boundary that define two visually interlocked regions in the surface. We have implemented this method by mapping appropriate textures to each face of the initial mesh. The resulting textured surfaces look aesthetically pleasing since they closely resemble planar TSP (traveling salesmen problem) art and Truchet-like curves. Graphical abstractDisplay Omitted Highlights? A method to convert any mesh surface into a single closed 3D curve is called surface covering. ? Two methods to construct 3D wires from surface covering curves. ? A method to obtain surface covering curves that can divide any given mesh surface into two regions.

The quality of the reconstructed image with 255 discrete levels from its halftoned version of homogenously distributed dot patterns depends on how well the fine textures can be represented and how the noisy dot patterns can be simultaneously removed on the flat regions. To satisfy these criteria, an iterative inverse halftoning method based on the texture-enhancing deconvolution with error-compensating feedback is presented. In this study, the input halftoned image is initially low-pass filtered with the Gaussian filtering to generate a blurred image on which the textures or details are sharply restored and the noisy halftoned dots on the flat regions are forced to be suppressed through the joint of the texture-enhancing deconvolution with spatially varying image priors and the structure-preserving image denoising. Moreover, the initially blurred image is iteratively updated with the addition of the created error image defined as the difference-image between the low-pass-filtered input halftoned image and the low-pass filtered halftoned image of the reconstructed continuous image, thereby compensating the missing textures. This error compensation is conducted until the stop criterion is satisfied. The experiment results showed that the proposed method not only reproduced the fine textures or details but also suppressed the noisy dots on the flat regions, more than the conventional state-of-the-art methods. Highlights? An input halftoned image is low-pass-filtered to create initially blurred image. ? A joint of the deconvolution and denoising is conducted to give the restored image. ? The initially blurred image is iteratively updated through an error feedback system. ? The restored image can be upgraded by enhancing fine texture and reducing noisy dots.