Curvature Information Research Articles

Accurate 3D Finger Knuckle Recognition Using Auto-Generated Similarity Functions

Contactless 3D finger knuckle is an emerging biometric identifier, which can provide a promising alternative for personal identification. To maximize its potential, feature representation and matching are the two critical components towards high performance. A recent pioneering work is limited by its preliminary design of the feature descriptor and the tailor-made similarity function. Although this method demonstrates decent recognition performance, there is room for improvement. This article advances the state-of-the-art method by introducing a new curvature based feature descriptor and a method to compute the similarity functions based on the statistical distribution of the encoded feature space. Our proposed feature representation utilizes an insight in 3D geometry for accurately encoding the curvature information. When computing the similarity between a pair of templates, we compute the similarity function from the probability mass distributions of the encoded feature space. Our proposed approach is scalable to templates with different sizes, and more importantly outperforms the state-of-the-art methods significantly, which is demonstrated in a publicly available database of 3D finger knuckle. In addition, we also demonstrate the generalizability of our approach by evaluating on other publicly available biometric datasets of similar patterns, i.e., 3D palmprint and finger vein.

Lean meat yield estimation using a prototype 3D imaging approach

Lean Meat Yield (LMY, %) of carcass is an important industry trait, which currently is not routinely measured in Australian beef abattoirs. Objective on-line technology to determine LMY is key for wider adoption. This paper presents a proof-of-concept approach for estimating the LMY of beef carcasses from the 3D information provided by RGB-D cameras. Moreover, a specifically designed on-line data acquisition system for abattoir applications is presented, consisting of three cameras moving on a scanning rig to generate 3D carcass side reconstructions. The hindquarter is then segmented consistently across all the 3D models to extract curvature information and LMY estimated via non-linear regression based on Gaussian Process models. Sides from 119 carcasses at two different commercial abattoirs were used to evaluate this approach. Results from this preliminary study (RMSE = 3.91%, R2 = 0.69) using curvature, P8 fat and HSCW indicate that 3D imaging of beef carcasses is a viable and relatively accurate technology to estimate LMY.

Color Image Retrieval Method Fusing Principal Curvature and Color Information

<p indent=0mm>Aiming at the problems of high feature dimensionality, long retrieval time, and low precision of existing algorithms in image retrieval system, an image retrieval algorithm based on the fusion of color information and texture information is proposed. Firstly, the color image is quantized and counted to obtain color histogram. Secondly, the Hessian matrix is used to extract curvature information of the color image, and the weighted joint histogram of oriented gradient based on curvature information is constructed. Then, the obtained color histogram and histogram of oriented gradient are fused to obtain the final joint histogram. And the experimental results on the Corel-1k, FTVL, and COIL-100 datasets show that the proposed algorithm has the advantages of high precision and low dimensionality.

Improved Strategy of Two‐Node Curved Beam Element Based on the Same Beam’s Nodes Information

The curved beam with a great initial curvature is the typical structure and applied widely in real engineering structures. The common practice in the current literature employs two‐node straight beam elements as the elementary members for stress and displacement analysis, which needs a large number of divisions to fit the curved beam shape well and increases computational time greatly. In this paper, we develop an improved accurate two‐node curved beam element (IC2) in 3D problems, combining the curved Timoshenko beam theory and the curvature information calculated from the same beam curve. The strategy of calculating the curvature information from the same bean curve in the IC2 beam element and then transferring the curvature information to the two‐node straight beam element can greatly enhance the accuracy of the mechanical analysis with no extra calculation burden. We then introduce the finite element implementation of the IC2 beam element and verify by the complex curved beam analysis. By comparison with simulation results from the straight two‐node beam element in the MIDAS (S2‐MIDAS) and the three‐node curved beam element adopted in the ANSYS (C3‐ANSYS), the simulation results of the typical quarter arc examples under constant or variable curvature show that the IC2 beam element based on curved beam theory is a combination of efficiency and accuracy. And, it is a good choice for analysis of complex engineering rod structure with large initial curvature.

Vehicle speed preview control with road curvature information for safety and comfort promotion

Vehicle lateral acceleration is a critical state and index for vehicle safety and ride comfort. To limit it in high speed cornering situation, a vehicle speed preview controller is proposed with the information of future road curvature, just as a human driver behavior. The future road curvature can be obtained from high definition map in intelligent vehicle control, and to implement it, model predictive control method (MPC) is implemented taking advantage of its preview nature. In this preview speed control framework, a novel kinematics model with vehicle location, speed and track curvature is established for vehicle states prediction. The control performance index of MPC is constructed with vehicle road following index and lateral acceleration index with the aiming of promoting safety and ride comfort. The controller is evaluated during cornering with different road trajectory, initial speed, preview time and road adhesion coefficient in a hardware-in-the-loop simulation platform. It is testified that vehicle slows down before cornering as human driver does to decrease lateral acceleration and steering angle with the benefit of promoting comfort and safety.

Generalized matrix transformation formalism for reflection and transmission of complex optical waves at a plane dielectric interface.

We describe a generalized formalism, addressing the fundamental problem of reflection and transmission of complex optical waves at a plane dielectric interface. Our formalism involves the application of generalized operator matrices to the incident constituent plane-wave fields to obtain the reflected and transmitted fields. This formalism, though physically equivalent to Fresnel formalism, has greater mathematical elegance and computational efficiency as compared to the latter. We utilize exact 3D electric-field expressions, which enable us to seamlessly analyze waves of miscellaneous wavefront shapes and properties using the single formalism, along with appropriately retaining the geometric phase and wavefront curvature information. We demonstrate our formalism by obtaining and analyzing the reflected and transmitted fields in a simulated Gaussian beam model.

Predictive Coupled Control of the Road Train Longitudinal and Lateral Motion on a Curved Path

In this paper, we consider the problem of a road train path-following on a curved path with an optimal velocity. To solve the problem, we propose a control algorithm based on the coupled model predictive control strategy. Model predictive control assumes the computation of a control sequence by solving an optimal control problem on a finite horizon for a current state of a nonlinear time-varying system. We use the truck steering angle and road train acceleration as control inputs. We describe the road train longitudinal and lateral dynamics using an implicit nonlinear model in continuous time. To derive a discrete linear time-varying state-space prediction model describing the deviations of system dynamics from a reference path we use the Euler method to discretize the original system and compute analytical formulae for its Jacobian by MATLAB Symbolic Math Toolbox. We calculate the reference path and corresponding reference values of the state vector applying the well-known geometric techniques, which utilize the path coordinates and its curvature information. We take the reference values of a truck and a semitrailer yaw angles to be equal. Thus, the reference value of the jackknifing angle is zero. The calculations of reference velocity take into account its skid and rollover limits. To validate the proposed path-following algorithm on the road train we design a simulation model in Simulink. The paper presents the simulation results of testing the movement of a road train along a given path for various values of the reference speed. We show that the algorithm provides high enough reference path-following accuracy, vehicle reference speed tracking, and low values of the jackknifing angle on the speed values up to 18 m/s and curvature radii down to 250 m. The proposed algorithm can be used in ADAS-systems and autonomous vehicles development.

A Modified Descent Spectral Conjugate Gradient Method for Unconstrained Optimization

Recently, Peyghami et al. (Optim Meth Soft 43:1–28, 2015) proposed a modified secant equation, which applied a new updating Yabe and Takano’s rule as an adaptive version of the parameter of conjugate gradient. Here, using this modified secant equation, we propose a new modified descent spectral nonlinear conjugate gradient method. The proposed method includes two major features; the higher-order accuracy in approximating the second-order curvature information of the objective function and the sufficient descent condition. Global convergence of the proposed method is proved for uniformly convex functions and general functions. Numerical experiments are done on a set of test functions of the CUTEr collection. The results are compared with some well-known methods.

A Stochastic Quasi-Newton Method for Large-Scale Nonconvex Optimization With Applications.

Ensuring the positive definiteness and avoiding ill conditioning of the Hessian update in the stochastic Broyden-Fletcher-Goldfarb-Shanno (BFGS) method are significant in solving nonconvex problems. This article proposes a novel stochastic version of a damped and regularized BFGS method for addressing the above problems. While the proposed regularized strategy helps to prevent the BFGS matrix from being close to singularity, the new damped parameter further ensures the positivity of the product of correction pairs. To alleviate the computational cost of the stochastic limited memory BFGS (LBFGS) updates and to improve its robustness, the curvature information is updated using the averaged iterate at spaced intervals. The effectiveness of the proposed method is evaluated through the logistic regression and Bayesian logistic regression problems in machine learning. Numerical experiments are conducted by using both synthetic data set and several real data sets. The results show that the proposed method generally outperforms the stochastic damped LBFGS (SdLBFGS) method. In particular, for problems with small sample sizes, our method has shown superior performance and is capable of mitigating ill-conditioned problems. Furthermore, our method is more robust to the variations of the batch size and memory size than the SdLBFGS method.

A Point Cloud Simplification Method Based on Modified Fuzzy C-Means Clustering Algorithm with Feature Information Reserved

To further improve the performance of the point cloud simplification algorithm and reserve the feature information of parts point cloud, a new method based on modified fuzzy c-means (MFCM) clustering algorithm with feature information reserved is proposed. Firstly, the normal vector, angle entropy, curvature, and density information of point cloud are calculated by combining principal component analysis (PCA) and k-nearest neighbors (k-NN) algorithm, respectively; Secondly, gravitational search algorithm (GSA) is introduced to optimize the initial cluster center of fuzzy c-means (FCM) clustering algorithm. Thirdly, the point cloud data combined coordinates with its feature information are divided by the MFCM algorithm. Finally, the point cloud is simplified according to point cloud feature information and simplified parameters. The point cloud test data are simplified using the new algorithm and traditional algorithms; then, the results are compared and discussed. The results show that the new proposed algorithm can not only effectively improve the precision of point cloud simplification but also reserve the accuracy of part features.

Horndeski gravity as D → 4 limit of Gauss-Bonnet

We propose a procedure for the D→4 limit of Einstein-Gauss-Bonnet (EGB) gravity that leads to a well defined action principle in four dimensions. Our construction is based on compactifying D-dimensional EGB gravity on a (D−4)-dimensional maximally symmetric space followed by redefining the Gauss-Bonnet coupling α→αD−4. The resulting model is a special scalar-tensor theory that belongs to the family of Horndeski gravity. Static black hole solutions in the scalar-tensor theory are investigated. Interestingly, the metric profile is independent of the curvature of the internal space and coincides with the D→4 limit of the usual EGB black hole with the unusual Gauss-Bonnet coupling αD−4. The curvature information of the internal space is instead encoded in the profile of the extra scalar field. Our procedure can also be generalized to define further limits of the Gauss-Bonnet combination by compactifying the D-dimensional theory on a (D−p)-dimensional maximally symmetric space with p≤3. These lead to different D→4 limits of EGB gravity as well as its D→2,3 limits.

Tactile line drawings for improved shape understanding in blind and visually impaired users

Members of the blind and visually impaired community rely heavily on tactile illustrations - raised line graphics on paper that are felt by hand - to understand geometric ideas in school textbooks, depict a story in children's books, or conceptualize exhibits in museums. However, these illustrations often fail to achieve their goals, in large part due to the lack of understanding in how 3D shapes can be represented in 2D projections. This paper describes a new technique to design tactile illustrations considering the needs of blind individuals. Successful illustration design of 3D objects presupposes identification and combination of important information in topology and geometry. We propose a twofold approach to improve shape understanding. First, we introduce a part-based multi-projection rendering strategy to display geometric information of 3D shapes, making use of canonical viewpoints and removing reliance on traditional perspective projections. Second, curvature information is extracted from cross sections and embedded as textures in our illustrations.

Fiber curvature sensor based on dual-wavelength erbium-doped fiber laser (DWEDFL) with gain competitive intensity difference demodulation

A fiber laser curvature sensor based on dual-wavelength gain competitive intensity difference demodulation is proposed and experimentally demonstrated. Here a linear cavity dual-wavelength erbium-doped fiber laser (DWEDFL) is build with two filter components i.e. fiber Bragg gratings (FBG) with very close center wavelengths. Since the dual wavelengths are sharing the same gain medium erbium-doped fiber (EDF), gain competition mechanism is formed with sensitive feature to external measurand curvature. Besides, the curvature information is demodulated by subtracting the power of the dual wavelengths. Experimental results show a total sensitivity as high as -57.22 dB/m−1. Our scheme not only improves the measuring sensitivity, but also eliminates the effects of source power jitter and external disturbances for example temperature shift since they do almost the same impact on the dual wavelengths.

A composite beam integrating an in-situ FPCB sensor membrane with PVDF arrays for modal curvature measurement

This research develops a novel composite beam integrating a flexible printed circuit board (FPCB) sensor membrane with polyvinylidene fluoride (PVDF) arrays for accurately extracting modal curvature in a convenient way, which can be easily used in the potential aerospace applications of in-situ structure damage detection or structure health monitoring. The proposed composite beam structure consists of a metal beam, a FPCB sensor membrane and a periodic array of identical rectangular PVDF films. Specifically, a periodic array of identical rectangular PVDF films is encapsulated in the FPCB sensor membrane, and the FPCB sensor membrane is integrated with the metal beam. Since the lead wire, sensors, the protective layer of sensor and the host beam structure are of one-piece construction in the proposed structure, the proposed structure has the self-sensing function that the discrete curvature of modal shapes of the metal beam can be directly obtained by using the experiment modal analysis (EMA) method and no other sensors. Results show that the proposed structure has good self-sensing performance and can directly extract the modal curvature information without any numerical differential calculation. In addition, the proposed structure restrains surrounding noise and improves the stability of the sensor array signal since the sensor array is encapsulated with the metal beam under the same condition.

3D Pose and Curvature Estimation of Bendable Interventional Device using Single-view X-ray Image.

We present a novel method to estimate the 3D pose and curvature of bendable interventional devices using a single X-ray image. A preliminary experiment was performed to demonstrate the feasibility of the proposed method. The mean estimation accuracies were 3.48mm and 0.59mm for the 3D pose and the radius of curvature of the bendable cardiac ablation catheter, respectively. This method has the potential to help clinicians to make a better intraoperative decision during the procedure, resulting in expedite surgery and reduce exposure to ionizing radiation.Clinical relevance- This novel method has the potential to improve clinician's intraoperative decision-making by providing the 3D pose and curvature information of bendable interventional devices, such as flexible catheter and endoscope.

Local eigenvalue decomposition for embedded Riemannian manifolds

Local Principal Component Analysis can be performed over small domains of an embedded Riemannian manifold in order to relate the covariance analysis of the underlying point set with the local extrinsic and intrinsic curvature. We show that the volume of domains on a submanifold of general codimension, determined by the intersection with higher-dimensional cylinders and balls in the ambient space, have asymptotic expansions in terms of the mean and scalar curvatures. Moreover, we propose a generalization of the classical third fundamental form to general submanifolds and prove that the local eigenvalue decomposition (EVD) of the covariance matrices have asymptotic expansions that contain the curvature information encoded by the traces of this tensor. This proves the general correspondence between the local EVD integral invariants and differential-geometric curvature for arbitrary embedded Riemannian submanifolds, found so far for curves and hypersurfaces only. Thus, we establish a key theoretical bridge, via covariance matrices at scale, for potential applications in manifold learning relating the statistics of point clouds sampled from Riemannian submanifolds to the underlying geometry.

Automatic pipe and elbow recognition from three-dimensional point cloud model of industrial plant piping system using convolutional neural network-based primitive classification

With the recent development of laser scanning technology, the variety of applications of laser scanners has increased. One typical application is object recognition from laser-scanned point cloud models. On large-scale construction sites such as refineries and industrial plants, object recognition from point cloud models has been widely employed for construction progress monitoring, assembly inspections, and maintenance purposes. Pipelines are among the main objects of interest with regard to object recognition on such sites. There has been extensive research on recognizing pipes in pipelines; however, research on recognizing pipe-connecting elbows is still lacking. Most representative elbow recognition methods are centerline-based and connectivity-based methods. These methods do not use laser-scanned points directly; instead, they employ feature values that are calculated from laser-scanned points. However, these feature values are easily affected by noise and occlusion; therefore, the elbow recognition results could be inaccurate owing to noisy and occluded point cloud models. In this paper, we propose an automatic pipe and elbow recognition method robust against noise and occlusion in which pipes and elbows are recognized directly from laser-scanned points. This method starts with pipeline extraction, followed by elbow classification based on curvature information. Falsely classified points are filtered using convolutional neural network-based primitive classification. After elbow recognition is completed, pipe classification and recognition are performed. Experimental results obtained from three different point cloud models demonstrated that the proposed method recognizes pipes and elbows with high accuracy from noisy and occluded point cloud models.

Utilization of LiDAR DTM for Systematic Improvement in Mapping and Classification of Coastal Micro-Geomorphology

The dynamicity of the coastal landforms necessitates monitoring from space-borne or aerial remote sensing platforms. Precise elevation information is an important requirement to delineate the subtle topographic variations along and across the coastal axes, indicative of coastal geomorphology at large scale. The study was carried out for a coastal stretch of 25 km, located between Visakhapatnam and Kakinada districts of Andhra Pradesh state, where the depositional activity is dominated. In the present study, we aim to demonstrate the utility of high resolution digital terrain model derived from the advanced Light Detection and Ranging (LiDAR) technology, in identification and delineation of the micro-geomorphic units along this coastline. The four level classifications of the landforms were also proposed based on genesis of landforms, geological sub-provinces, landform shape and morphological attributes as level 1, 2, 3 and 4, respectively. The higher-order landforms of level 4, such as beach berms, beach ridge crest, dune crest and toe, interdunal depression, etc. are delineated using the cross-shore elevation profile, slope and signed curvature information derived from LiDAR DTM with the subtle contrasts of elevation (1–2 m). The beach ridge–swale complex, dune and beach berms were chosen as the key depositional elements to define the morphological attributes in terms of shape and curvature values. This has enabled creation of micro-geomorphology map at 1:5000 scales over the study area as well as classification of the landforms into 4 levels, thus providing significant input for coastal resource mapping and management.

Accelerated full-waveform inversion using dynamic mini-batches

SUMMARYWe present an accelerated full-waveform inversion based on dynamic mini-batch optimization, which naturally exploits redundancies in observed data from different sources. The method rests on the selection of quasi-random subsets (mini-batches) of sources, used to approximate the misfit and the gradient of the complete data set. The size of the mini-batch is dynamically controlled by the desired quality of the gradient approximation. Within each mini-batch, redundancy is minimized by selecting sources with the largest angular differences between their respective gradients, and spatial coverage is maximized by selecting candidate events with Mitchell’s best-candidate algorithm. Information from sources not included in a specific mini-batch is incorporated into each gradient calculation through a quasi-Newton approximation of the Hessian, and a consistent misfit measure is achieved through the inclusion of a control group of sources. By design, the dynamic mini-batch approach has several main advantages: (1) The use of mini-batches with adaptive size ensures that an optimally small number of sources is used in each iteration, thus potentially leading to significant computational savings; (2) curvature information is accumulated and exploited during the inversion, using a randomized quasi-Newton method; (3) new data can be incorporated without the need to re-invert the complete data set, thereby enabling an evolutionary mode of full-waveform inversion. We illustrate our method using synthetic and real-data inversions for upper-mantle structure beneath the African Plate. In these specific examples, the dynamic mini-batch approach requires around 20 per cent of the computational resources in order to achieve data and model misfits that are comparable to those achieved by a standard full-waveform inversion where all sources are used in each iteration.

Directional bending sensor based on triangular shaped fiber Bragg gratings.

We propose a directional bending sensor by mechanically bonded FBGs in a radially placed triangular shape to measure the bending moment and its applied direction. Bending measurement and the determination of its direction is essential to get the maximum bending stress and to figure out the deflection shape by the curvature information in a wide range of engineering applications. The proposed sensor configuration is original in terms of a pointwise sensing scheme based on the mechanically bonded scheme of three optical fibers and its simple fabrication process only using optical fibers. This set-up made of three FBG sensors provides a simple relationship of the bending moment to the independent strain measurement of each FBG. We demonstrate two sensor configurations such as the self-bending measurement and the structural bending measurement in a surface-mount (or embedded) measurement. Through the verification experiments, the proposed bending sensor is successfully shown to measure the applied bending moment and its direction.