Thin-plate Spline Research Articles

Spatial Distribution of Multi-Fractal Scaling Behaviours of Atmospheric XCO2 Concentration Time Series during 2010-2018 over China.

Exploring the spatial distribution of the multi-fractal scaling behaviours in atmospheric CO2 concentration time series is useful for understanding the dynamic mechanisms of carbon emission and absorption. In this work, we utilise a well-established multi-fractal detrended fluctuation analysis to examine the multi-fractal scaling behaviour of a column-averaged dry-air mole fraction of carbon dioxide (XCO2) concentration time series over China, and portray the spatial distribution of the multi-fractal scaling behaviour. As XCO2 data values from the Greenhouse Gases Observing Satellite (GOSAT) are insufficient, a spatio-temporal thin plate spline interpolation method is applied. The results show that XCO2 concentration records over almost all of China exhibit a multi-fractal nature. Two types of multi-fractal sources are detected. One is long-range correlations, and the other is both long-range correlations and a broad probability density function; these are mainly distributed in southern and northern China, respectively. The atmospheric temperature and carbon emission/absorption are two possible external factors influencing the multi-fractality of the atmospheric XCO2 concentration. Highlight: (1) An XCO2 concentration interpolation is conducted using a spatio-temporal thin plate spline method. (2) The spatial distribution of the multi-fractality of XCO2 concentration over China is shown. (3) Multi-fractal sources and two external factors affecting multi-fractality are analysed.

An ontogenetic study of Eospirigerina putilla (Brachiopoda) surviving the Late Ordovician mass extinction

The ontogeny of the brachiopod can be reflected by the shell outline, size, and their interconnections. As a representative early atrypid brachiopod group of the Edgewood-Cathay Fauna, the genus Eospirigerina survived the Late Ordovician mass extinction (LOME). Abundant E. putilla specimens were collected from the “Wulipo Bed” limestone (upper Hirnantian, Ordovician). As a typical Eospirigerina dominated Edgewood-Cathay Fauna from the upper Hirnantian (Ordovician) of northeastern Yunnan, China, the quantitative ontogenetic study of E. putilla can provide a better understanding of its palaeoecology. Using the image contour recognition, we develop an automated measuring tool, test its robustness, and measure 554 complete and silicified specimens of E. putilla. An abnormal palaeoenvironment that the E. putilla population inhabited in is suggested by the high mortality of small individuals indicated by the size-frequency histogram and survivorship curve. Using quadratic polynomial regression, the negative allometric pattern of the shell length to width in the population is recognized. Using geometric morphometrics with semi-landmarks for 51 specimens from the population, the ontogenic trends are mapped by PCA with thin-plate splines, in which the heterochronic shifts may occur from an elongate shell with relatively larger ventral cardinal area and delthyrium to wider outline and smaller beak. Such heterochronic development is possibly caused by the unstable and stressful environment immediately after the LOME.

Landmark-Based Geometric Morphometric of Apis dorsata and A. d. binghami Wing Venation in Indonesian Archipelagos

The giant honey bee Apis dorsata has a vast distribution in Asia, including in the Indonesian archipelagos, whereas A. d. binghami is endemic to Sulawesi Island. This research aimed to analyze the variations of landmarks in wing venation of A. dorsata from Sumatra, Belitung, Java, Sumbawa, Moluccas islands, and A. d. binghami in Sulawesi, based on geometric morphometric analysis. Nineteen landmarks from 200 wing venations were analyzed using Thin Plate Spline software. Our study found landmarks 16 and 17, known as Cubital Index in traditional morphometrics, contributed to the wing shape of A. dorsata and A. d. binghami Among all samples, these two landmarks show high displacement in the wing shape of A. d. binghami in Sulawesi and A. dorsata in the Moluccas. On the contrary, we found that the Sumatra, Belitung, and Java Apis dorsata revealed low displacement in the deformation grid; resulting in their shapes being more similar to the reference landmark. The variations of wing shape separate A. d. binghami from Sulawesi and Sumbawa in the Principle Component Analysis and agreement with the Neighbor-joining tree. Therefore, the geometric morphometric based on landmarks of wing venation is a powerful tool to discriminate the subspecies level of A. dorsata.

Terrain sensitive climate mapping for the Arequipa Department in Peru

AbstractClimate is a powerful driver of agricultural and natural systems, and spatial climate datasets are currently in great demand. This is especially true in the Arequipa Department of Peru, a region with low seasonal precipitation, remarkable topographic variability, and significant water demand in a highly managed water system. This paper presents the Arequipa Climate Maps (ACM) datasets, a high resolution (1 km) spatial 30‐year (1988–2017) climate dataset for the Arequipa Region, in Peru. Four interpolation methods, and combinations of those methods, were tested to produce 30 years of daily precipitation, maximum and minimum air temperature: Ordinary Kriging (OK), Thin Plate Splines (TPS), Regression Kriging (RK), and Regression Thin Plate Splines (RTPS). The mixed method RTPS‐TPS and RTPS using locally fitted polynomial and potential regressions were found to best represent the spatial variability of precipitation and daily extreme temperatures, respectively, and helped compensate the bias resulting from the lack of weather stations at higher elevations. These methods were then selected to create the ACM dataset, which contains climate maps of 30‐year annual and monthly climate normals (ACM–Normals) and 30 years of annual, monthly, and daily climate maps (ACM–YMD). In addition, insights on weather station gap filling in mountainous areas and bias corrections for avoidance of anomalous precipitation and to assure consistency between annual, monthly and daily data are presented, together with discussion about the quality and limitations of the dataset, and its comparison with other datasets.

An improved non-stationary geostatistical method for three-dimensional interpolation of Benzo(a)pyrene at a contaminated site

Intense industrial activities and complex hydrogeological conditions at contaminated sites make accurate three-dimensional (3D) mapping challenging. The cause is the non-stationarity in the variance of soil pollutants in geographical space (G-space), making the stationary hypothesis required by the Kriging method unsatisfactory. To handle the variance non-stationarity, a Variance-Octree-Kriging (VOK) method was proposed. VOK is a spatial deformation method that constructs a stationary deformation space (D-space) by stretching and shrinking the G-spaces with low and high spatial correlation, respectively. VOK method consists of 3D stratification in G-space, space scaling and transformation, and ordinary Kriging (OK) in D-space. 3D stratification uses variance octree (VOT) to generate a set of anchor points in the G-space. The spatial scaling and transformation use the virtual force algorithm (VFA) and thin-plate spline to evenly distribute the anchor points and obtain the D-space, where the OK is implemented. The method was applied to predict the distribution of soil Benzo(a)pyrene (BaP) at a contaminated site in North China Plain. The results show that the interpolation accuracy of VOK was 9% higher than that of OK. The VOK method also changed the spatial structure from anisotropic to isotropic. The root mean squared error (RMSE) of fill, silt and clay layers decreased by 4.67%, 11.39%, and 20.46%, respectively. This method is applicable to the 3D interpolation of pollutants at contaminated sites, with the advantages of high interpolation accuracy and the ability to handle the non-stationarity in variance.

Free vibration analysis of laminated rectangular plates with varying thickness using Legendre-radial point interpolation method

The purpose of this paper is to investigate the free vibration characteristics of laminated rectangular plate with varying thickness (LRPVT) using the meshfree strong form method. The theoretical formulation for free vibration analysis is based on the first order shear deformation theory (FSDT), and the governing equations for the LRPVT are derived using Hamilton's principle. In order to approximate the displacement function in the governing equations, a novel Legendre-radial point interpolation method (LRPIM) shape function using combined basis of thin plate spline (TPS) radial functions and Legendre polynomials is presented. The boundary conditions are generalized by the introduction of an artificial spring technique, and the type of boundary conditions is selected according to the spring stiffness values. The accuracy and reliability of the LRPIM shape function are verified by comparing with the results of the literatures and finite element software ABAQUS. The free vibration characteristics including natural frequencies and mode shapes of the LRPVT with various geometry and boundary conditions are presented through numerical examples.

Geometric morphometric analysis of the pharyngeal airway during treatment of Class III malocclusion

The pharyngeal airway is important during diagnosis and treatment planning in orthodontics. This study aimed to compare the changes in the shape of the pharyngeal airway in patients with Class III malocclusion treated with a facemask (FM) with a removable maxillary appliance and modified tandem traction bow appliance (MTTBA). This study consisted of pretreatment and posttreatment lateral cephalograms of 35 subjects with skeletal and dental Class III malocclusion. In the first group, 20 patients (12 males, 8 females; mean age, 10.2 years) were treated with MTTBA. The mean treatment time was 12 months. In the second group, 15 patients (10 males, 5 females; mean age, 10.3 years) were treated with FM. The mean treatment time was 11.7 months. Nineteen pharyngeal landmarks were considered from the image corresponding to the midsagittal plane and marked using tpsDig software (version 2.04; Stony Brook, NY). Pharyngeal airway shape difference between the groups was assessed by performing a Generalized Procrustes analysis. The shape deformation of the pharyngeal airway from the pre- to posttreatment periods was evaluated using the thin-plate spline method. There were no differences between MTTBA and FM groups according to airway shape for pretreatment and posttreatment periods. However, there were some deformities using the enlargement of the nasopharyngeal area in the FM group and oropharyngeal area in the MTTBA group according to pretreatment periods. There were no differences between the groups according to the mean pharyngeal airway shapes when the posttreatment periods of the FM and MTTBA groups were examined.

Geodesics in the TPS Space

In shape analysis, the interpolation of shapes’ trajectories is often performed by means of geodesics in an appropriate Riemannian Shape Space. Over the past several decades, different metrics and shape spaces have been proposed, including Kendall shape space, LDDMM based approaches, and elastic contour, among others. Once a Riemannian space is chosen, geodesics and parallel transports can be used to build splines or piecewise geodesics paths. In a recent paper, we introduced a new Riemannian shape space named TPS Space based on the Thin Plate Spline interpolant and characterized by an appropriate metric and parallel transport rule. In the present paper, we further explore the geometry of the TPS Space by characterizing the properties of its geodesics. Several applications show the capability of the proposed formulation to conserve important physical properties of deformation, such as local strains and global elastic energy.

A Knuckle‐Walking Signal in the Trabecular Architecture of the African Ape Scaphoid

African ape (Panand Gorilla) carpals have few diagnostic features in their external morphology reflecting knuckle‐walking hand postures, and whether the internal trabecular architecture of their carpals yields a functional signal for knuckle‐walking is unclear. Here we perform a high‐density, multiple‐VOI analysis of trabecular architecture in the scaphoid of knuckle‐walking African apes and non‐knuckle‐walkers (Pongo, Macaca,and Papio). We predict that (1) Gorilla, the taxon with largest body mass, will have the highest trabecular thickness (Tb.Th) and trabecular spacing (Tb.Sp) because Tb.Th and Tb.Sp have been shown to be correlated with body size, (2) bone volume fraction (BV/TV) will be higher in habitually terrestrial taxa than in arboreal taxa, as the forelimb is thought to experience relatively high compressive loads during terrestrial locomotion, (3) Pongo will have the lowest degree of anisotropy (DA) as they engage in variable locomotor behaviors, and (4) pattern of DA distribution will differ between knuckle‐walking and non‐knuckle‐walking taxa, as the consistency and direction of wrist joint loading likely differs among primates with different quadrupedal hand postures.Fifty‐four scaphoids were scanned using high‐resolution µCT (voxel range: 48‐27 μm). Surface models were created from image data and then subjected to a weighted spherical harmonic analysis. Volumes of interest (VOIs) were manually placed into a surface model of the average scaphoid, and then warped to individual specimens using the thin plate spline interpolation function. Trabecular properties were calculated using the BoneJ plugin for ImageJ and MATLAB, and variation in distribution of trabecular parameters among taxa was summarized using a principal component analysis (PCA). Parameter values among species were compared with non‐parametric MANOVAs using a randomized residual permutation procedure, and PC scores were compared using Kruskal‐Wallis ANOVAs. The parameter distribution within each bone was visualized by comparing species‐average color maps of average parameter values.As predicted, Tb.Th and Tb.Sp are greatest in Gorilla, indicating that these parameters are related in part to body size. In contrast, results indicate that BV/TV and DA magnitudes do not follow predictions. Specifically, habitually terrestrial taxa do not have significantly higher BV/TV than arboreal taxa, and taxa that engage in more variable locomotor behaviors (i.e., Pongo) do not have the lowest values of DA. Finally, results indicate that knuckle‐walking African apes can be distinguished from other taxa studied here by pattern of DA distribution, and these differences appear to reflect differences in habitual hand postures and locomotor modes. If this pattern is unique to knuckle‐walking apes among hominoids, and is also identified in the scaphoid of fossil hominoids, then it may be used to indicate whether fossil taxa were habitual knuckle‐walkers.

Phenotypic variation of Chitala chitala (Hamilton, 1822) from Indian rivers using truss network and geometric morphometrics.

Chitala chitala (Hamilton, 1822) is an economically important food fish species occurring throughout Indian rivers, which also has ornamental value. This study focuses on morphological variations in C. chitala from seven river basins across India namely; Son, Tons, Ken, Brahmaputra, Ganga, Gomti and Gandak. A truss network was constructed by interconnecting nine landmarks to generate 36 morphometric variables extracted from digital images of specimens sampled from the study locations. Transformed truss measurements were subjected to principal component analysis (PCA), canonical discriminant function analysis (CDFA) and discriminant analyses of principal components (DAPC). DAPC function coefficients performed much better in capturing the variation pattern and discrimination between the rivers which was not achieved using CDFA. Eight truss variables were identified with significant and highest loading for truss variables on principal components and coefficients on discriminant function from DAPC contributing to maximum variation between the rivers. Performance graph and functional distribution of identified truss variables clearly indicated distinction between the rivers. Thin plate spline analysis and procrustes shape analysis further showed the variation in morphology between specimens across the rivers. The significant parameters differentiating specimens from different rivers were linked to dorsal fin origin, the base of the pectoral fin and the perpendicular point on the anal fin from the dorsal fin origin. Variation in the hydrodynamics of the rivers studied might be possibly affecting the fin kinematics and consequently leading to adaption seen as phenotypic variation in C. chitala. The results showcased in the present study shall help in better understanding of intra-specific diversity which is significant for management and conservation of a species.

Multi-manifold discriminant local spline embedding

Manifold learning reveals the intrinsic low-dimensional manifold structure of high-dimensional data and has achieved great success in a wide spectrum of applications. However, traditional manifold learning methods assume that all the data lie on a common manifold, hence fail to capture the complicated geometry structure of the real-world data lying on multiple manifolds. This paper proposes a novel Multi-manifold Discriminant Local Spline Embedding (MDLSE) algorithm for high-dimensional classification, which considers a more realistic scenario where data of the same class lies on the same manifold. On the basis of this assumption, MDLSE seeks to reconstruct multiple manifolds for different classes of data in the embedding and separate them as apart as possible. In order to preserve the geometry structure of all the manifolds, MDLSE employs thin plate splines to align the local patches within each manifold compatibly in the global embedding. Meanwhile, to separate the different manifolds, MDLSE utilizes discriminative information to ensure the neighboring data from different manifolds to be mapped far from each other. Extensive experiments on both synthetic and real-world datasets demonstrate the superiority of MDLSE over the other representative manifold learning algorithms. The advantage of MDLSE is often more obvious on smaller size of training data and in lower embedding dimensions.

The association of indoor heat exposure with diabetes and respiratory 9-1-1 calls through emergency medical dispatch and services documentation

BackgroundPeople with pre-existing medical conditions, who spend a large proportion of their time indoors, are at risk of emergent morbidities from elevated indoor heat exposures. In this study, indoor heat of structures wherein exposed people received Grady Emergency Services based care in Atlanta, GA, U.S., was measured from May to September 2016. Methodology: In this case-control study, analyses were conducted to investigate the effect of indoor heat on the odds of 9-1-1 calls for diabetic (n = 90 cases) and separately, for respiratory (n = 126 cases), conditions versus heat-insensitive emergencies (n = 698 controls). Generalized Additive Models considered both linear and non-linear indoor heat and health outcome associations using thin-plate regression splines. ResultsHotter and more humid indoor conditions were non-linearly associated with an increasing likelihood of receiving emergency care for complications of diabetes and severe respiratory distress. Higher heat indices were associated with increased odds of a diabetes (odds ratio for change from 30 to 31 °C: 1.12, 95% CI: 1.08–1.16) or respiratory 9-1-1 medical call versus control (odds ratio for change from 34 to 35 °C: 1.18, 95% CI: 1.09–1.28) call. Both diabetic and respiratory distress patients were more likely to be African-American and/or have comorbidities. ConclusionsIn this study, the statistical association of indoor heat exposure with emergency morbidities (diabetic, respiratory) was demonstrated. The study also showcased the value and utility of data gathered by emergency medical dispatch and services from inaccessible private indoor sources (i.e., domiciles) for environmental health.

A Comparative Study of Different Schemes Based on Bézier-like Functions with an Application of Craniofacial Fractures Reconstruction

Cranial implants, especially custom made implants, are complex, important and necessary in craniofacial fracture restoration surgery. However, the classical procedure of the manual design of the implant is time consuming and complicated. Different computer-based techniques proposed by different researchers, including CAD/CAM, mirroring, reference skull, thin plate spline and radial basis functions have been used for cranial implant restoration. Computer Aided Geometric Design (CAGD) has also been used in bio-modeling and specifically for the restoration of cranial defects in form of different spline curves, namely C1,C2,GC1GC2, rational curves, B-spline and Non-Uniform Rational B-Spline (NURBS) curves. This paper gives an in-depth comparison of existing techniques by highlighting the limitations and advantage in different contexts. The construction of craniofacial fractures is made using different Bézier-like functions (Ball, Bernstein and Timmer basis functions) and is analyzed in detail. The C1,GC1 and GC2 cubic Ball curves are performed well for construction of the small fractured part. Any form of fracture is constructed using this approach and it has been effectively applied to frontal and parietal bone fractures. However, B-spline and NURBS curves can be used for any type of fractured parts and are more friendly user.

Qualitative and quantitative analyses of the Carboniferous ostracod genus Carbonita from Rod El-Hamal area, Wadi Araba, north Eastern Desert, Egypt

ABSTRACT The present work provides a qualitative analysis of the studied ostracod carapaces of the genus Carbonita. Thirty-seven ostracod specimens were collected from the Late Carboniferous Rod El-Hamal Formation at Rod El-Hamal area, Wadi Araba, north Eastern Desert, Egypt. The collected ostracods were assigned to four species in genus Carbonita; C. fabulina (Jones and Kirkby), C. inflata (Jones and Kirkby), C. humilis (Jones and Kirkby), and C. bairdoides (Jones and Kirkby) according to the qualitative traits. In order to assess the discrimination among the different Carbonita species, a new technique, geometric morphometric analyses, was carried out using ImageJ and MorphoJ software packages. Geometric morphometric analyses included landmark configurations, procrustes superimposition, principle component analysis, and thin plate spline deformation plots. Using nine landmarks, average shapes of the different Carbonita species have been provided. The principle component analysis separated Carbonita fabulina and C. bairdoides from the morphologically overlapped C. inflata and C. humilis. Shape variations in the carapace length, height, anteroventral arc, posterodorsal arc, posteroventral arc, and the anterior margin were demonstrated through the deformation plots. These quantitative analyses, along with the qualitative examination, provided considerable variations among the different Carbonita species, and hence the best way for species discrimination.

Radial basis function-based vector field algorithm for wildfire boundary tracking with UAVs

This paper tackles the problem of dynamic wildfire boundary tracking with UAVs. Wildfire boundary is treated as the zero-level set curve of an implicit function and is approximated with radial basis functions. Its propagation is modeled with the Hamilton–Jacobi equation with an arbitrary initial boundary as the input. To navigate UAVs to the wildfire boundary, an analytical velocity vector field, whose integral curves converge to the wildfire boundary, is constructed on the basis of the typical radial basis function thin-plate spline. Computer simulations with a single UAV and multiple UAVs have been conducted for the evaluation of the proposed solution, and numerical results show that the proposed algorithm can ensure the successful tracking of an arbitrarily shaped wildfire boundary.

Spatially varying coefficient models using reduced-rank thin-plate splines

Spatially varying coefficient (SVC) regression models are concerned about regression for spatial data, where regression coefficients may vary in space. This paper proposes a new approach for SVC modeling by representing regression coefficients using a class of multiresolution spline basis functions in a generalized-linear model framework. The proposed method provides flexible and parsimonious representations for regression coefficients. It enables commonly used (generalized) linear-regression packages for estimation, testing, and constructing confidence levels. We develop a fast estimation algorithm that simultaneously performs variable selection, detects spatial heterogeneity for each variable, and determines its complexity. We provide numerical examples and an application to a real estate dataset to demonstrate the proposed method’s effectiveness.

Spatial+: A novel approach to spatial confounding.

In spatial regression models, collinearity between covariates and spatial effects can lead to significant bias in effect estimates. This problem, known as spatial confounding, is encountered modeling forestry data to assess the effect of temperature on tree health. Reliable inference is difficult as results depend on whether or not spatial effects are included in the model. We propose a novel approach, spatial+, for dealing with spatial confounding when the covariate of interest is spatially dependent but not fully determined by spatial location. Using a thin plate spline model formulation we see that, in this case, the bias in covariate effect estimates is a direct result of spatial smoothing. Spatial+ reduces the sensitivity of the estimates to smoothing by replacing the covariates by their residuals after spatial dependence has been regressed away. Through asymptotic analysis we show that spatial+ avoids the bias problems of the spatial model. This is also demonstrated in a simulation study. Spatial+ is straightforward to implement using existing software and, as the response variable is the same as that of the spatial model, standard model selection criteria can be used for comparisons. A major advantage of the method is also that it extends to models with non-Gaussian response distributions. Finally, while our results are derived in a thin plate spline setting, the spatial+ methodology transfers easily to other spatial modelformulations.

Spatial analysis of seasonal precipitation using various interpolation methods in the Euphrates basin, Turkey

It is vital to accurately map the spatial distribution of precipitation, which is widely used in many fields such as hydrology, climatology, meteorology, ecology, and agriculture. This study aimed to reveal the spatial distribution of seasonal, long-term average precipitation in the Euphrates Basin with various interpolation methods. For this reason, Simple Kriging, Ordinary Kriging, Universal Kriging, Ordinary CoKriging, Empirical Bayesian Kriging, Radial Basis Functions (Completely Regularized Spline, Thin Plate Spline, Multiquadratic, Inverse Multiquadratic, Spline with Tensor), Local Polynomial Interpolation, Global Polynomial Interpolation, and Inverse Distance Weighting methods have been applied in the Geographical Information Systems environment. Long-term seasonal precipitation averages between 1966 and 2017 are presented as input for predicting precipitation maps. The accuracy of the precipitation prediction maps was based on linear regression analysis, root mean square error (RMSE), mean absolute error (MAE), correlation coefficient (R), and determination coefficient (R2) values obtained from the cross-validation tests. The most suitable method was chosen for the interpolation method that gives the lowest RMSE, MAE, and the largest R and R2. As a result of the study, Ordinary CoKriging in spring and winter precipitation, Local Polynomial Interpolation in summer precipitation, and Ordinary Kriging in autumn precipitation were the most appropriate estimation methods.

The potential and limits of Thin-Plate Spline retrodeformation on asymmetrical objects:simulation of taphonomic deformations andapplication on a fossil sample of limb long bones

In this study, we suggest a method adapted to the retrodeformation of asymmetrical objects – such as limb bones – by quantitatively estimating the effectiveness of the Thin-Plate Splines (TPS) interpolation function as a retrodeformation tool. To do so, taphonomic deformations were first simulated on a single horse femur. The original bone was then used as a reference in order to drive the retrodeformation using anatomical landmarks. This approach, based on a single bone, enabled us to evaluate the performance of the retrodeformation procedure. Then, the same approach was performed on a sample of rhino femora but using a different specimen (from the same species) as the reference in order to account for morphological variation. We also added sliding semi-landmarks on anatomical curves. Finally, retrodeformation was applied on a sample of sauropodomorph dinosaur femora by building a mean shape based on several well-preserved fossil specimens. Results show that entirely flattened and stretched bones are more efficiently retrodeformed than bent and twisted bones. Introduction of morphological variation increased the efficiency of retrodeformation for bent and locally stretched bones. The application to the sample of fossils produced similar results but also highlighted the difficulty of retrodeforming bones with a combination of different deformations. TPS interpolation is an efficient tool of retrodeformation for asymmetrical objects, especially for bones with only one affine deformation such as flattening or stretching. Finding a threshold of landmark number to use for this process would be the next step because it would allow us to ensure the quality of retrodeformation while keeping available a reasonable number of landmarks in order to perform shape analysis on retrodeformed bones. Twisted and bent fossils are frequently discovered and we suggest that these kinds of deformations should be studied with caution, especially when combined with other types of taphonomic distortions.

User-oriented Natural Human-Robot Control with Thin-Plate Splines and LRCN

We propose a real-time vision-based teleoperation approach for robotic arms that employs a single depth-based camera, exempting the user from the need for any wearable devices. By employing a natural user interface, this novel approach leverages the conventional fine-tuning control, turning it into a direct body pose capture process. The proposed approach is comprised of two main parts. The first is a nonlinear customizable pose mapping based on Thin-Plate Splines (TPS), to directly transfer human body motion to robotic arm motion in a nonlinear fashion, thus allowing matching dissimilar bodies with different workspace shapes and kinematic constraints. The second is a Deep Neural Network hand-state classifier based on Long-term Recurrent Convolutional Networks (LRCN) that exploits the temporal coherence of the acquired depth data. We validate, evaluate and compare our approach through both classical cross-validation experiments of the proposed hand state classifier; and user studies over a set of practical experiments involving variants of pick-and-place and manufacturing tasks. Results revealed that LRCN networks outperform single image Convolutional Neural Networks; and that users' learning curves were steep, thus allowing the successful completion of the proposed tasks. When compared to a previous approach, the TPS approach revealed no increase in task complexity and similar times of completion, while providing more precise operation in regions closer to workspace boundaries.