An Efficient Multiple Object Detection and Tracking Framework for Automatic Counting and Video Surveillance Applications

Automatic visual object counting and video surveillance have important applications for home and business environments, such as security and management of access points. However, in order to obtain a satisfactory performance these technologies need professional and expensive hardware, complex installations and setups, and the supervision of qualified workers. In this paper, an efficient visual detection and tracking framework is proposed for the tasks of object counting and surveillance, which meets the requirements of the consumer electronics: off-the-shelf equipment, easy installation and configuration, and unsupervised working conditions. This is accomplished by a novel Bayesian tracking model that can manage multimodal distributions without explicitly computing the association between tracked objects and detections. In addition, it is robust to erroneous, distorted and missing detections. The proposed algorithm is compared with a recent work, also focused on consumer electronics, proving its superior performance.

Efficient methodologies for system matrix modelling in iterative image reconstruction for rotating high-resolution PET

A fully 3D iterative image reconstruction algorithm has been developed for high-resolution PET cameras composed of pixelated scintillator crystal arrays and rotating planar detectors, based on the ordered subsets approach. The associated system matrix is precalculated with Monte Carlo methods that incorporate physical effects not included in analytical models, such as positron range effects and interaction of the incident gammas with the scintillator material. Custom Monte Carlo methodologies have been developed and optimized for modelling of system matrices for fast iterative image reconstruction adapted to specific scanner geometries, without redundant calculations. According to the methodology proposed here, only one-eighth of the voxels within two central transaxial slices need to be modelled in detail. The rest of the system matrix elements can be obtained with the aid of axial symmetries and redundancies, as well as in-plane symmetries within transaxial slices. Sparse matrix techniques for the non-zero system matrix elements are employed, allowing for fast execution of the image reconstruction process. This 3D image reconstruction scheme has been compared in terms of image quality to a 2D fast implementation of the OSEM algorithm combined with Fourier rebinning approaches. This work confirms the superiority of fully 3D OSEM in terms of spatial resolution, contrast recovery and noise reduction as compared to conventional 2D approaches based on rebinning schemes. At the same time it demonstrates that fully 3D methodologies can be efficiently applied to the image reconstruction problem for high-resolution rotational PET cameras by applying accurate pre-calculated system models and taking advantage of the system's symmetries.

Efficient management of backtracking in and-parallelism

A backtracking algorithm for AND-Parallelism and its implementation at the Abstract Machine level are presented: first, a class of AND-Parallelism models based on goal independence is defined, and a generalized version of Restricted AND-Parallelism (RAP) introduced as characteristic of this class. A simple and efficient backtracking algorithm for R A P is then discussed. An implementation scheme is presented for this algorithm which offers minimum overhead, while retaining the performance and storage economy of sequent ial implementations and taking advantage of goal independence to avoid unnecessary backtracking ("restricted intelligent backtracking"). Finally, the implementation of backtracking in sequential and AND-Parallcl systems is explained through a number of examples.

Mesh traversal and sorting for efficient memory usage in scientific codes

Applications that operate on meshes are very popular in High Performance Computing (HPC) environments. In the past, many techniques have been developed in order to optimize the memory accesses for these datasets. Different loop transformations and domain decompositions are com- monly used for structured meshes. However, unstructured grids are more challenging. The memory accesses, based on the mesh connectivity, do not map well to the usual lin- ear memory model. This work presents a method to improve the memory performance which is suitable for HPC codes that operate on meshes. We develop a method to adjust the sequence in which the data are used inside the algorithm, by means of traversing and sorting the mesh. This sorted mesh can be transferred sequentially to the lower memory levels and allows for minimum data transfer requirements. The method also reduces the lower memory requirements dra- matically: up to 63% of the L1 cache misses are removed in a traditional cache system. We have obtained speedups of up to 2.58 on memory operations as measured in a general- purpose CPU. An improvement is also observed with se- quential access memories, where we have observed reduc- tions of up to 99% in the required low-level memory size.

Efficient random variable generation: ratio of uniforms and polar rejection sampling

Monte Carlo techniques, which require the generation of samples from some target density, are often the only alternative for performing Bayesian inference. Two classic sampling techniques to draw independent samples are the ratio of uniforms (RoU) and rejection sampling (RS). An efficient sampling algorithm is proposed combining the RoU and polar RS (i.e. RS inside a sector of a circle using polar coordinates). Its efficiency is shown in drawing samples from truncated Cauchy and Gaussian random variables, which have many important applications in signal processing and communications. RESUMEN. Método eficiente para generar algunas variables aleatorias de uso común en procesado de señal y comunicaciones (por ejemplo, Gaussianas o Cauchy truncadas) mediante la combinación de dos técnicas: "ratio of uniforms" y "rejection sampling".

Efficient adaptive methods based on adjoint equations and truncation error estimation for unstructured grids

El propósito de esta tesis es la implementación de métodos eficientes de adaptación de mallas basados en ecuaciones adjuntas en el marco de discretizaciones de volúmenes finitos para mallas no estructuradas. La metodología basada en ecuaciones adjuntas optimiza la malla refinándola adecuadamente con el objetivo de mejorar la precisión de cálculo de un funcional de salida dado. El funcional suele ser una magnitud escalar de interés ingenieril obtenida por post-proceso de la solución, como por ejemplo, la resistencia o la sustentación aerodinámica. Usualmente, el método de adaptación adjunta está basado en una estimación a posteriori del error del funcional de salida mediante un promediado del residuo numérico con las variables adjuntas, “Dual Weighted Residual method” (DWR). Estas variables se obtienen de la solución del problema adjunto para el funcional seleccionado. El procedimiento habitual para introducir este método en códigos basados en discretizaciones de volúmenes finitos involucra la utilización de una malla auxiliar embebida obtenida por refinamiento uniforme de la malla inicial. El uso de esta malla implica un aumento significativo de los recursos computacionales (por ejemplo, en casos 3D el aumento de memoria requerida respecto a la que necesita el problema fluido inicial puede llegar a ser de un orden de magnitud). En esta tesis se propone un método alternativo basado en reformular la estimación del error del funcional en una malla auxiliar más basta y utilizar una técnica de estimación del error de truncación, denominada _ -estimation, para estimar los residuos que intervienen en el método DWR. Utilizando esta estimación del error se diseña un algoritmo de adaptación de mallas que conserva los ingredientes básicos de la adaptación adjunta estándar pero con un coste computacional asociado sensiblemente menor. La metodología de adaptación adjunta estándar y la propuesta en la tesis han sido introducidas en un código de volúmenes finitos utilizado habitualmente en la industria aeronáutica Europea. Se ha investigado la influencia de distintos parámetros numéricos que intervienen en el algoritmo. Finalmente, el método propuesto se compara con otras metodologías de adaptación de mallas y su eficiencia computacional se demuestra en una serie de casos representativos de interés aeronáutico. ABSTRACT The purpose of this thesis is the implementation of efficient grid adaptation methods based on the adjoint equations within the framework of finite volume methods (FVM) for unstructured grid solvers. The adjoint-based methodology aims at adapting grids to improve the accuracy of a functional output of interest, as for example, the aerodynamic drag or lift. The adjoint methodology is based on the a posteriori functional error estimation using the adjoint/dual-weighted residual method (DWR). In this method the error in a functional output can be directly related to local residual errors of the primal solution through the adjoint variables. These variables are obtained by solving the corresponding adjoint problem for the chosen functional. The common approach to introduce the DWR method within the FVM framework involves the use of an auxiliary embedded grid. The storage of this mesh demands high computational resources, i.e. over one order of magnitude increase in memory relative to the initial problem for 3D cases. In this thesis, an alternative methodology for adapting the grid is proposed. Specifically, the DWR approach for error estimation is re-formulated on a coarser mesh level using the _ -estimation method to approximate the truncation error. Then, an output-based adaptive algorithm is designed in such way that the basic ingredients of the standard adjoint method are retained but the computational cost is significantly reduced. The standard and the new proposed adjoint-based adaptive methodologies have been incorporated into a flow solver commonly used in the EU aeronautical industry. The influence of different numerical settings has been investigated. The proposed method has been compared against different grid adaptation approaches and the computational efficiency of the new method has been demonstrated on some representative aeronautical test cases.

Efficient speed advisories for multi-stage-metering arrival management

This paper presents a methodology and algorithm for Air Traffic Control (ATC) to efficiently achieve schedules arrival times through speed control in the presence of uncertainty. The methodology does not assume the availability of airborne time of arrival control and can therefore be applied to legacy aircraft. The speed advisories are calculated in a manner that allows for sufficient control margin to, if required, adjust the aircraft's trajectory at a later stage to correct for estimated arrival time drift at the lowest impact to efficiency. The methodology is therefore envisioned to prevent major last-minute interventions and instead assists ATC in allowing more continuous descent approaches to be conducted by aircraft leading to more efficient operations.

Fully adaptive gaussian mixture metropolis-hastings algorithm

Markov Chain Monte Carlo methods are widely used in signal processing and communications for statistical inference and stochastic optimization. In this work, we introduce an efficient adaptive Metropolis-Hastings algorithm to draw samples from generic multimodal and multidimensional target distributions. The proposal density is a mixture of Gaussian densities with all parameters (weights, mean vectors and covariance matrices) updated using all the previously generated samples applying simple recursive rules. Numerical results for the one and two-dimensional cases are provided.

Efficient Monte Carlo optimization for multi-label classifier chains

Multi-label classification (MLC) is the supervised learning problem where an instance may be associated with multiple labels. Modeling dependencies between labels allows MLC methods to improve their performance at the expense of an increased computational cost. In this paper we focus on the classifier chains (CC) approach for modeling dependencies. On the one hand, the original CC algorithm makes a greedy approximation, and is fast but tends to propagate errors down the chain. On the other hand, a recent Bayes-optimal method improves the performance, but is computationally intractable in practice. Here we present a novel double-Monte Carlo scheme (M2CC), both for finding a good chain sequence and performing efficient inference. The M2CC algorithm remains tractable for high-dimensional data sets and obtains the best overall accuracy, as shown on several real data sets with input dimension as high as 1449 and up to 103 labels.

Feasibility assessment of the interactive use of a Monte Carlo algorithm in treatment planning for intraoperative electron radiation therapy

This work analysed the feasibility of using a fast, customized Monte Carlo (MC) method to perform accurate computation of dose distributions during pre- and intraplanning of intraoperative electron radiation therapy (IOERT) procedures. The MC method that was implemented, which has been integrated into a specific innovative simulation and planning tool, is able to simulate the fate of thousands of particles per second, and it was the aim of this work to determine the level of interactivity that could be achieved. The planning workflow enabled calibration of the imaging and treatment equipment, as well as manipulation of the surgical frame and insertion of the protection shields around the organs at risk and other beam modifiers. In this way, the multidisciplinary team involved in IOERT has all the tools necessary to perform complex MC dosage simulations adapted to their equipment in an efficient and transparent way. To assess the accuracy and reliability of this MC technique, dose distributions for a monoenergetic source were compared with those obtained using a general-purpose software package used widely in medical physics applications. Once accuracy of the underlying simulator was confirmed, a clinical accelerator was modelled and experimental measurements in water were conducted. A comparison was made with the output from the simulator to identify the conditions under which accurate dose estimations could be obtained in less than 3 min, which is the threshold imposed to allow for interactive use of the tool in treatment planning. Finally, a clinically relevant scenario, namely early-stage breast cancer treatment, was simulated with pre- and intraoperative volumes to verify that it was feasible to use the MC tool intraoperatively and to adjust dose delivery based on the simulation output, without compromising accuracy. The workflow provided a satisfactory model of the treatment head and the imaging system, enabling proper configuration of the treatment planning system and providing good accuracy in the dosage simulation.

Efficient Monte Carlo methods for multi-dimensional learning with classifier chains

Multi-dimensional classification (MDC) is the supervised learning problem where an instance is associated with multiple classes, rather than with a single class, as in traditional classification problems. Since these classes are often strongly correlated, modeling the dependencies between them allows MDC methods to improve their performance – at the expense of an increased computational cost. In this paper we focus on the classifier chains (CC) approach for modeling dependencies, one of the most popular and highest-performing methods for multi-label classification (MLC), a particular case of MDC which involves only binary classes (i.e., labels). The original CC algorithm makes a greedy approximation, and is fast but tends to propagate errors along the chain. Here we present novel Monte Carlo schemes, both for finding a good chain sequence and performing efficient inference. Our algorithms remain tractable for high-dimensional data sets and obtain the best predictive performance across several real data sets.

Logarithmical hopping encoding: a low computational complexity algorithm for image compression

LHE (logarithmical hopping encoding) is a computationally efficient image compression algorithm that exploits the Weber–Fechner law to encode the error between colour component predictions and the actual value of such components. More concretely, for each pixel, luminance and chrominance predictions are calculated as a function of the surrounding pixels and then the error between the predictions and the actual values are logarithmically quantised. The main advantage of LHE is that although it is capable of achieving a low-bit rate encoding with high quality results in terms of peak signal-to-noise ratio (PSNR) and image quality metrics with full-reference (FSIM) and non-reference (blind/referenceless image spatial quality evaluator), its time complexity is O( n) and its memory complexity is O(1). Furthermore, an enhanced version of the algorithm is proposed, where the output codes provided by the logarithmical quantiser are used in a pre-processing stage to estimate the perceptual relevance of the image blocks. This allows the algorithm to downsample the blocks with low perceptual relevance, thus improving the compression rate. The performance of LHE is especially remarkable when the bit per pixel rate is low, showing much better quality, in terms of PSNR and FSIM, than JPEG and slightly lower quality than JPEG-2000 but being more computationally efficient.

Deciduous tree reconstruction algorithm based on cylinder fitting from mobile terrestrial laser scanned point clouds

Vector reconstruction of objects from an unstructured point cloud obtained with a LiDAR-based system (light detection and ranging) is one of the most promising methods to build three dimensional models of orchards. The cylinder fitting method for woody structure reconstruction of leafless trees from point clouds obtained with a mobile terrestrial laser scanner (MTLS) has been analysed. The advantage of this method is that it performs reconstruction in a single step. The most time consuming part of the algorithm is generation of the cylinder direction, which must be recalculated at the inclusion of each point in the cylinder. The tree skeleton is obtained at the same time as the cluster of cylinders is formed. The method does not guarantee a unique convergence and the reconstruction parameter values must be carefully chosen. A balanced processing of clusters has also been defined which has proven to be very efficient in terms of processing time by following the hierarchy of branches, predecessors and successors. The algorithm was applied to simulated MTLS of virtual orchard models and to MTLS data of real orchards. The constraints applied in the method have been reviewed to ensure better convergence and simpler use of parameters. The results obtained show a correct reconstruction of the woody structure of the trees and the algorithm runs in linear logarithmic time

Efficient tool path computation using multi-core GPUs

Tool path generation is one of the most complex problems in Computer Aided Manufacturing. Although some efficient strategies have been developed, most of them are only useful for standard machining. However, the algorithms used for tool path computation demand a higher computation performance, which makes the implementation on many existing systems very slow or even impractical. Hardware acceleration is an incremental solution that can be cleanly added to these systems while keeping everything else intact. It is completely transparent to the user. The cost is much lower and the development time is much shorter than replacing the computers by faster ones. This paper presents an optimisation that uses a specific graphic hardware approach using the power of multi-core Graphic Processing Units (GPUs) in order to improve the tool path computation. This improvement is applied on a highly accurate and robust tool path generation algorithm. The paper presents, as a case of study, a fully implemented algorithm used for turning lathe machining of shoe lasts. A comparative study will show the gain achieved in terms of total computing time. The execution time is almost two orders of magnitude faster than modern PCs.

QRT: A Qr-based tridiagonalization algorithm for nonsymmetric matrices

The stable similarity reduction of a nonsymmetric square matrix to tridiagonal form has been a long-standing problem in numerical linear algebra. The biorthogonal Lanczos process is in principle a candidate method for this task, but in practice it is confined to sparse matrices and is restarted periodically because roundoff errors affect its three-term recurrence scheme and degrade the biorthogonality after a few steps. This adds to its vulnerability to serious breakdowns or near-breakdowns, the handling of which involves recovery strategies such as the look-ahead technique, which needs a careful implementation to produce a block-tridiagonal form with unpredictable block sizes. Other candidate methods, geared generally towards full matrices, rely on elementary similarity transformations that are prone to numerical instabilities. Such concomitant difficulties have hampered finding a satisfactory solution to the problem for either sparse or full matrices. This study focuses primarily on full matrices. After outlining earlier tridiagonalization algorithms from within a general framework, we present a new elimination technique combining orthogonal similarity transformations that are stable. We also discuss heuristics to circumvent breakdowns. Applications of this study include eigenvalue calculation and the approximation of matrix functions.