Automatic modeling of pectus excavatum corrective prosthesis using artificial neural networks

Pectus excavatum is the most common deformity of the thorax. Pre-operative diagnosis usually includes Computed Tomography (CT) to successfully employ a thoracic prosthesis for anterior chest wall remodeling. Aiming at the elimination of radiation exposure, this paper presents a novel methodology for the replacement of CT by a 3D laser scanner (radiation-free) for prosthesis modeling. The complete elimination of CT is based on an accurate determination of ribs position and prosthesis placement region through skin surface points. The developed solution resorts to a normalized and combined outcome of an artificial neural network (ANN) set. Each ANN model was trained with data vectors from 165 male patients and using soft tissue thicknesses (STT) comprising information from the skin and rib cage (automatically determined by image processing algorithms). Tests revealed that ribs position for prosthesis placement and modeling can be estimated with an average error of 5.0 ± 3.6 mm. One also showed that the ANN performance can be improved by introducing a manually determined initial STT value in the ANN normalization procedure (average error of 2.82 ± 0.76 mm). Such error range is well below current prosthesis manual modeling (approximately 11 mm), which can provide a valuable and radiation-free procedure for prosthesis personalization.

Automatic modeling of pectus excavatum corrective prosthesis using artificial neural networks

Pectus excavatum is the most common deformity of the thorax. Pre-operative diagnosis usually includes Computed Tomography (CT) to successfully employ a thoracic prosthesis for anterior chest wall remodeling. Aiming at the elimination of radiation exposure, this paper presents a novel methodology for the replacement of CT by a 3D laser scanner (radiation-free) for prosthesis modeling. The complete elimination of CT is based on an accurate determination of ribs position and prosthesis placement region through skin surface points. The developed solution resorts to a normalized and combined outcome of an artificial neural network (ANN) set. Each ANN model was trained with data vectors from 165 male patients and using soft tissue thicknesses (STT) comprising information from the skin and rib cage (automatically determined by image processing algorithms). Tests revealed that ribs position for prosthesis placement and modeling can be estimated with an average error of 5.0 ± 3.6 mm. One also showed that the ANN performance can be improved by introducing a manually determined initial STT value in the ANN normalization procedure (average error of 2.82 ± 0.76 mm). Such error range is well below current prosthesis manual modeling (approximately 11 mm), which can provide a valuable and radiation-free procedure for prosthesis personalization.

Neural-network approach to modeling liquid crystals in complex confinement

Finding the structure of a confined liquid crystal is a difficult task since both the density and order parameter profiles are nonuniform. Starting from a microscopic model and density-functional theory, one has to either (i) solve a nonlinear, integral Euler-Lagrange equation, or (ii) perform a direct multidimensional free energy minimization. The traditional implementations of both approaches are computationally expensive and plagued with convergence problems. Here, as an alternative, we introduce an unsupervised variant of the multilayer perceptron (MLP) artificial neural network for minimizing the free energy of a fluid of hard nonspherical particles confined between planar substrates of variable penetrability. We then test our algorithm by comparing its results for the structure (density-orientation profiles) and equilibrium free energy with those obtained by standard iterative solution of the Euler-Lagrange equations and with Monte Carlo simulation results. Very good agreement is found and the MLP method proves competitively fast, flexible, and refinable. Furthermore, it can be readily generalized to the richer experimental patterned-substrate geometries that are now experimentally realizable but very problematic to conventional theoretical treatments.

Neural-network approach to modeling liquid crystals in complex confinement

Finding the structure of a confined liquid crystal is a difficult task since both the density and order parameter profiles are nonuniform. Starting from a microscopic model and density-functional theory, one has to either (i) solve a nonlinear, integral Euler-Lagrange equation, or (ii) perform a direct multidimensional free energy minimization. The traditional implementations of both approaches are computationally expensive and plagued with convergence problems. Here, as an alternative, we introduce an unsupervised variant of the multilayer perceptron (MLP) artificial neural network for minimizing the free energy of a fluid of hard nonspherical particles confined between planar substrates of variable penetrability. We then test our algorithm by comparing its results for the structure (density-orientation profiles) and equilibrium free energy with those obtained by standard iterative solution of the Euler-Lagrange equations and with Monte Carlo simulation results. Very good agreement is found and the MLP method proves competitively fast, flexible, and refinable. Furthermore, it can be readily generalized to the richer experimental patterned-substrate geometries that are now experimentally realizable but very problematic to conventional theoretical treatments.

Computational modeling of resting-state activity demonstrates markers of normalcy in children with prenatal or perinatal stroke.

Children who sustain a prenatal or perinatal brain injury in the form of a stroke develop remarkably normal cognitive functions in certain areas, with a particular strength in language skills. A dominant explanation for this is that brain regions from the contralesional hemisphere "take over" their functions, whereas the damaged areas and other ipsilesional regions play much less of a role. However, it is difficult to tease apart whether changes in neural activity after early brain injury are due to damage caused by the lesion or by processes related to postinjury reorganization. We sought to differentiate between these two causes by investigating the functional connectivity (FC) of brain areas during the resting state in human children with early brain injury using a computational model. We simulated a large-scale network consisting of realistic models of local brain areas coupled through anatomical connectivity information of healthy and injured participants. We then compared the resulting simulated FC values of healthy and injured participants with the empirical ones. We found that the empirical connectivity values, especially of the damaged areas, correlated better with simulated values of a healthy brain than those of an injured brain. This result indicates that the structural damage caused by an early brain injury is unlikely to have an adverse and sustained impact on the functional connections, albeit during the resting state, of damaged areas. Therefore, these areas could continue to play a role in the development of near-normal function in certain domains such as language in these children.

HYBRID ARTIFICIAL NEURAL NETWORK APPLIEDTO MODELING SCFE OF BASIL AND ROSEMARY OILS

This work presents the results of a Hybrid Neural Network (HNN) technique as applied to modeling SCFE curves obtained from two Brazilian vegetable matrices. A series Hybrid Neural Network was employed to estimate the parameters of the phenomenological model. A small set of SCFE data of each vegetable was used to generate an extended data set, sufficient to train the network. Afterwards, other sets of experimental data, not used in the network training, were used to validate the present approach. The series HNN correlates well the experimental data and it is shown that the predictions accomplished with this technique may be promising for SCFE purposes.

Artificial neural network modeling of phenol adsorption onto barley husks activated carbon in an airlift reactor.

Tesis (Doctor en Ciencias con Orientación en Procesos Sustentables) UANL, 2013.

Modeling High-Dimensional Audio Sequences with Recurrent Neural Networks

Cette thèse étudie des modèles de séquences de haute dimension basés sur des réseaux de neurones récurrents (RNN) et leur application à la musique et à la parole. Bien qu'en principe les RNN puissent représenter les dépendances à long terme et la dynamique temporelle complexe propres aux séquences d'intérêt comme la vidéo, l'audio et la langue naturelle, ceux-ci n'ont pas été utilisés à leur plein potentiel depuis leur introduction par Rumelhart et al. (1986a) en raison de la difficulté de les entraîner efficacement par descente de gradient. Récemment, l'application fructueuse de l'optimisation Hessian-free et d'autres techniques d'entraînement avancées ont entraîné la recrudescence de leur utilisation dans plusieurs systèmes de l'état de l'art. Le travail de cette thèse prend part à ce développement. L'idée centrale consiste à exploiter la flexibilité des RNN pour apprendre une description probabiliste de séquences de symboles, c'est-à-dire une information de haut niveau associée aux signaux observés, qui en retour pourra servir d'à priori pour améliorer la précision de la recherche d'information. Par exemple, en modélisant l'évolution de groupes de notes dans la musique polyphonique, d'accords dans une progression harmonique, de phonèmes dans un énoncé oral ou encore de sources individuelles dans un mélange audio, nous pouvons améliorer significativement les méthodes de transcription polyphonique, de reconnaissance d'accords, de reconnaissance de la parole et de séparation de sources audio respectivement. L'application pratique de nos modèles à ces tâches est détaillée dans les quatre derniers articles présentés dans cette thèse. Dans le premier article, nous remplaçons la couche de sortie d'un RNN par des machines de Boltzmann restreintes conditionnelles pour décrire des distributions de sortie multimodales beaucoup plus riches. Dans le deuxième article, nous évaluons et proposons des méthodes avancées pour entraîner les RNN. Dans les quatre derniers articles, nous examinons différentes façons de combiner nos modèles symboliques à des réseaux profonds et à la factorisation matricielle non-négative, notamment par des produits d'experts, des architectures entrée/sortie et des cadres génératifs généralisant les modèles de Markov cachés. Nous proposons et analysons également des méthodes d'inférence efficaces pour ces modèles, telles la recherche vorace chronologique, la recherche en faisceau à haute dimension, la recherche en faisceau élagué et la descente de gradient. Finalement, nous abordons les questions de l'étiquette biaisée, du maître imposant, du lissage temporel, de la régularisation et du pré-entraînement.

Artificial neural network modeling for surface roughness prediction in cylindrical grinding of Al‐SiCp metal matrix composites and ANOVA analysis

Metal matrix composites (MMC) having aluminium (Al) in the matrix phase and silicon carbide particles (SiCp) in reinforcement phase, ie Al‐SiCp type MMC, have gained popularity in the re‐cent past. In this competitive age, manufacturing industries strive to produce superior quality products at reasonable price. This is possible by achieving higher productivity while performing machining at optimum combinations of process variables. The low weight and high strength MMC are found suitable for variety of components

Modeling of complex-valued Wiener systems using B-spline neural network

In this brief, a new complex-valued B-spline neural network is introduced in order to model the complex-valued Wiener system using observational input/output data. The complex-valued nonlinear static function in the Wiener system is represented using the tensor product from two univariate B-spline neural networks, using the real and imaginary parts of the system input. Following the use of a simple least squares parameter initialization scheme, the Gauss-Newton algorithm is applied for the parameter estimation, which incorporates the De Boor algorithm, including both the B-spline curve and the first-order derivatives recursion. Numerical examples, including a nonlinear high-power amplifier model in communication systems, are used to demonstrate the efficacy of the proposed approaches.

Modeling electrocortical activity through improved local approximations of integral neural field equations

Neural field models of firing rate activity typically take the form of integral equations with space-dependent axonal delays. Under natural assumptions on the synaptic connectivity we show how one can derive an equivalent partial differential equation (PDE) model that properly treats the axonal delay terms of the integral formulation. Our analysis avoids the so-called long-wavelength approximation that has previously been used to formulate PDE models for neural activity in two spatial dimensions. Direct numerical simulations of this PDE model show instabilities of the homogeneous steady state that are in full agreement with a Turing instability analysis of the original integral model. We discuss the benefits of such a local model and its usefulness in modeling electrocortical activity. In particular, we are able to treat “patchy” connections, whereby a homogeneous and isotropic system is modulated in a spatially periodic fashion. In this case the emergence of a “lattice-directed” traveling wave predicted by a linear instability analysis is confirmed by the numerical simulation of an appropriate set of coupled PDEs.

Complex-valued B-spline neural networks for modeling and inverting Hammerstein systems

Many communication signal processing applications involve modelling and inverting complex-valued (CV) Hammerstein systems. We develops a new CV B-spline neural network approach for efficient identification of the CV Hammerstein system and effective inversion of the estimated CV Hammerstein model. Specifically, the CV nonlinear static function in the Hammerstein system is represented using the tensor product from two univariate B-spline neural networks. An efficient alternating least squares estimation method is adopted for identifying the CV linear dynamic model’s coefficients and the CV B-spline neural network’s weights, which yields the closed-form solutions for both the linear dynamic model’s coefficients and the B-spline neural network’s weights, and this estimation process is guaranteed to converge very fast to a unique minimum solution. Furthermore, an accurate inversion of the CV Hammerstein system can readily be obtained using the estimated model. In particular, the inversion of the CV nonlinear static function in the Hammerstein system can be calculated effectively using a Gaussian-Newton algorithm, which naturally incorporates the efficient De Boor algorithm with both the B-spline curve and first order derivative recursions. The effectiveness of our approach is demonstrated using the application to equalisation of Hammerstein channels.