Multi-sensor background subtraction by fusing multiple region-based probabilistic classifiers

In the recent years, the computer vision community has shown great interest on depth-based applications thanks to the performance and flexibility of the new generation of RGB-D imagery. In this paper, we present an efficient background subtraction algorithm based on the fusion of multiple region-based classifiers that processes depth and color data provided by RGB-D cameras. Foreground objects are detected by combining a region-based foreground prediction (based on depth data) with different background models (based on a Mixture of Gaussian algorithm) providing color and depth descriptions of the scene at pixel and region level. The information given by these modules is fused in a mixture of experts fashion to improve the foreground detection accuracy. The main contributions of the paper are the region-based models of both background and foreground, built from the depth and color data. The obtained results using different database sequences demonstrate that the proposed approach leads to a higher detection accuracy with respect to existing state-of-the-art techniques.

Vacant parking area estimation through background subtraction and transience map analysis

A method for estimating the dimensions of non-delimited free parking areas by using a static surveillance camera is proposed. The proposed method is specially designed to tackle the main challenges of urban scenarios (multiple moving objects, outdoor illumination conditions and occlusions between vehicles) with no training. The core of this work is the temporal analysis of the video frames to detect the occupancy variation of the parking areas. Two techniques are combined: background subtraction using a mixture of Gaussians to detect and track vehicles and the creation of a transience map to detect the parking and leaving of vehicles. The authors demonstrate that the proposed method yields satisfactory estimates on three real scenarios while being a low computational cost solution that can be applied in any kind of parking area covered by a single camera.

Adaptive Multi-Camera System for Real Time Object Detection

In this paper we present an adaptive multi-camera system for real time object detection able to efficiently adjust the computational requirements of video processing blocks to the available processing power and the activity of the scene. The system is based on a two level adaptation strategy that works at local and at global level. Object detection is based on a Gaussian mixtures model background subtraction algorithm. Results show that the system can efficiently adapt the algorithm parameters without a significant loss in the detection accuracy.

A region based approach to background modeling in a wavelet multi-resolution framework

In the field of detection and monitoring of dynamic objects in quasi-static scenes, background subtraction techniques where background is modeled at pixel-level, although showing very significant limitations, are extensively used. In this work we propose a novel approach to background modeling that operates at region-level in a wavelet based multi-resolution framework. Based on a segmentation of the background, characterization is made for each region independently as a mixture of K Gaussian modes, considering the model of the approximation and detail coefficients at the different wavelet decomposition levels. Background region characterization is updated along time, and the detection of elements of interest is carried out computing the distance between background region models and those of each incoming image in the sequence. The inclusion of the context in the modeling scheme through each region characterization makes the model robust, being able to support not only gradual illumination and long-term changes, but also sudden illumination changes and the presence of strong shadows in the scene

Seasonal evolution of soil respiration in a mixed forestry plantation of walnuts (Juglans × intermedia Carr.) and alders (Alnus cordata (Loisel.) Duby) in Domaine de Restinclières (France)

The ecological intensification of crops is proposed as a solution to the growing demand of agricultural and forest resources, in opposition to intensive monocultures. The introduction of mixed cultures as mixtures between nitrogen fixing species and non nitrogen fixing species intended to increase crop yield as a result of an improvement of the available nitrogen and phosphorus in soil. Relationship between crops have received little attention despite the wide range of advantages that confers species diversity to these systems, such as increased productivity, resilience to disruption and ecological sustainability. Forests and forestry plantations can develop an important role in storing carbon in their tissues, especially in wood which become into durable product. A simplifying parameter to analyze the amount allocated carbon by plantation is the TBCA (total belowground carbon allocation), whereby, for short periods and mature plantations, is admitted as the subtraction between soil carbon efflux and litterfall. Soil respiration depends on a wide range of factors, such as soil temperature and soil water content, soil fertility, presence and type of vegetation, among others. The studied orchard is a mixed forestry plantation of hybrid walnuts(Juglans × intermedia Carr.) for wood and alders (Alnus cordata (Loisel.) Duby.), a nitrogen fixing specie through the actinomycete Frankia alni ((Woronin, 1866) Von Tubeuf 1895). The study area is sited at Restinclières, a green area near Montpellier (South of France). In the present work, soil respiration varied greatly throughout the year, mainly influenced by soil temperature. Soil water content did not significantly influence the response of soil respiration as it was constant during the measurement period and under no water stress conditions. Distance between nearest walnut and measurement was also a highly influential factor in soil respiration. Generally there was a decreasing trend in soil respiration when the distance to the nearest tree increased. It was also analyzed the response of soil respiration according to alder presence and fertilizer management (50 kg N·ha-1·año-1 from 1999 to 2010). None of these treatments significantly influenced soil respiration, although previous studies noticed an inhibition in rates of soil respiration under fertilized conditions and high rates of available nitrogen. However, treatments without fertilization and without alder presence obtained higher respiration rates in those cases with significant differences. The lack of significant differences between treatments may be due to the high coefficient of variation experienced by soil respiration measurements. Finally an asynchronous fluctuation was observed between soil respiration and litterfall during senescence period. This is possibly due to the slowdown in the emission of exudates by roots during senescence period, which are largely related to microbial activity.

Optoelectronic Properties of InAlN/GaN Distributed Bragg Reflector Heterostructure Examined by Valence

High-resolution monochromated electron energy loss spectroscopy (EELS) at subnanometric spatial resolution and <200 meV energy resolution has been used to assess the valence band properties of a distributed Bragg reflector multilayer heterostructure composed of InAlN lattice matched to GaN. This work thoroughly presents the collection of methods and computational tools put together for this task. Among these are zero-loss-peak subtraction and nonlinear fitting tools, and theoretical modeling of the electron scattering distribution. EELS analysis allows retrieval of a great amount of information: indium concentration in the InAlN layers is monitored through the local plasmon energy position and calculated using a bowing parameter version of Vegard Law. Also a dielectric characterization of the InAlN and GaN layers has been performed through Kramers-Kronig analysis of the Valence-EELS data, allowing band gap energy to be measured and an insight on the polytypism of the GaN layers.

FocusDET, a new toolbox for SISCOM analysis. Evaluation of the registration accuracy using monte carlo simulation

Subtraction of Ictal SPECT Co-registered to MRI (SISCOM) is an imaging technique used to localize the epileptogenic focus in patients with intractable partial epilepsy. The aim of this study was to determine the accuracy of registration algorithms involved in SISCOM analysis using FocusDET, a new user-friendly application. To this end, Monte Carlo simulation was employed to generate realistic SPECT studies. Simulated sinograms were reconstructed by using the Filtered BackProjection (FBP) algorithm and an Ordered Subsets Expectation Maximization (OSEM) reconstruction method that included compensation for all degradations. Registration errors in SPECT-SPECT and SPECT-MRI registration were evaluated by comparing the theoretical and actual transforms. Patient studies with well-localized epilepsy were also included in the registration assessment. Global registration errors including SPECT-SPECT and SPECT-MRI registration errors were less than 1.2 mm on average, exceeding the voxel size (3.32 mm) of SPECT studies in no case. Although images reconstructed using OSEM led to lower registration errors than images reconstructed with FBP, differences after using OSEM or FBP in reconstruction were less than 0.2 mm on average. This indicates that correction for degradations does not play a major role in the SISCOM process, thereby facilitating the application of the methodology in centers where OSEM is not implemented with correction of all degradations. These findings together with those obtained by clinicians from patients via MRI, interictal and ictal SPECT and video-EEG, show that FocusDET is a robust application for performing SISCOM analysis in clinical practice.

Foreground segmentation in depth imagery using depth and spatial dynamic models for video surveillance applications

Low-cost systems that can obtain a high-quality foreground segmentation almostindependently of the existing illumination conditions for indoor environments are verydesirable, especially for security and surveillance applications. In this paper, a novelforeground segmentation algorithm that uses only a Kinect depth sensor is proposedto satisfy the aforementioned system characteristics. This is achieved by combininga mixture of Gaussians-based background subtraction algorithm with a new Bayesiannetwork that robustly predicts the foreground/background regions between consecutivetime steps. The Bayesian network explicitly exploits the intrinsic characteristics ofthe depth data by means of two dynamic models that estimate the spatial and depthevolution of the foreground/background regions. The most remarkable contribution is thedepth-based dynamic model that predicts the changes in the foreground depth distributionbetween consecutive time steps. This is a key difference with regard to visible imagery,where the color/gray distribution of the foreground is typically assumed to be constant.Experiments carried out on two different depth-based databases demonstrate that theproposed combination of algorithms is able to obtain a more accurate segmentation of theforeground/background than other state-of-the art approaches.

Background foreground segmentation with RGB-D Kinect data: An efficient combination of classifiers

Low cost RGB-D cameras such as the Microsoft’s Kinect or the Asus’s Xtion Pro are completely changing the computer vision world, as they are being successfully used in several applications and research areas. Depth data are particularly attractive and suitable for applications based on moving objects detection through foreground/background segmentation approaches; the RGB-D applications proposed in literature employ, in general, state of the art foreground/background segmentation techniques based on the depth information without taking into account the color information. The novel approach that we propose is based on a combination of classifiers that allows improving background subtraction accuracy with respect to state of the art algorithms by jointly considering color and depth data. In particular, the combination of classifiers is based on a weighted average that allows to adaptively modifying the support of each classifier in the ensemble by considering foreground detections in the previous frames and the depth and color edges. In this way, it is possible to reduce false detections due to critical issues that can not be tackled by the individual classifiers such as: shadows and illumination changes, color and depth camouflage, moved background objects and noisy depth measurements. Moreover, we propose, for the best of the author’s knowledge, the first publicly available RGB-D benchmark dataset with hand-labeled ground truth of several challenging scenarios to test background/foreground segmentation algorithms.

Piel artificial : metamorfosis arquitectónica del cuerpo a través de la superficie

El cuerpo, como conjunto organizado de partes que configuran el organismo, es una entidad metamórfica. El ser humano procura dar continuidad a esta condición mutante que le caracteriza, mediante diversas acciones de carácter arquitectónico. A partir de la observación de los procesos naturales, el individuo se autodefine artificialmente, transformando su realidad innata en una versión distorsionada de sí misma. Por adición, sustracción o modificación, la piel como última capa natural, se convierte en lienzo de manipulación plástica primordial para asegurar la existencia y controlar la identidad, individual y colectiva. La evolución experimental de estas intervenciones primarias, permite suplantar la piel natural por una reinterpretación construida; una piel exenta y desmontable con la que proyectar un yo diferente provisionalmente. El uso constante de esta prótesis removible e intercambiable, provoca que el cuerpo desnudo se transforme en un cuerpo vestido, en un entorno social en el que la desnudez deja de ser el estado natural del ser humano. La piel artificial se construye mediante una gran diversidad de procesos proyectuales, siendo la transformación de la superficie bidimensional en envolvente tridimensional el más utilizado a lo largo de la existencia de la vestimenta. El plano, concebido como principal formato de revestimiento humano, se adapta a su irregularidad topográfica por modelado, perforación, fragmentación, trazado, parametrización e interacción, transformándose en una envolvente cada vez más compleja y perfecta. Su diseño implica la consideración de variables como la dimensión y la escala, la función y la forma, la estructura, el material y la construcción, la técnica y los instrumentos. La vestimenta es una arquitectura habitacional individual, un límite corporal que relaciona el espacio entre el exterior e el interior, lo ajeno y lo propio, el tú y el yo; un filtro concreto y abstracto simultáneamente; una interfaz en donde el vestido es el continente y el cuerpo su contenido. ABSTRACT The body as a whole, organized of parts that make up the organism, is a metamorphic entity. The human being seeks to give continuity to this mutant condition which characterizes him through various actions of architectural character. From the observation of the natural processes, the individual defines itself artificially, transforming its innate reality into a distorted version of itself. By addition, subtraction or modification, the skin, as the last natural layer, becomes canvas of primary plastic handling in order to ensure the existence and to control the identity, both individual and collective. The experimental evolution of these primary interventions allows to impersonate the natural skin by a constructed reinterpretation; a free and detachable skin together with which to be able to project, temporarily, a different “I”. The constant use of this removable and interchangeable prosthesis causes the naked body to be transformed into a dressed body, in a social setting in which the nudity is no longer the natural state of the human being. The artificial skin is constructed by a variety of projectual processes; the most used throughout the existence of the outfit is transforming the two-dimensional surface into a three-dimensional covering. The plan, conceived as the main human lining format, adapts to its topographic irregularity by modeling, drilling, fragmentation, outline, parameters and interaction, thus becoming a type of increasingly more complex and perfect covering. Its design implies the consideration of different variables such as the dimension and the scale, the function and the shape, the structure, the material and the construction, the technique and the instruments. The clothing is an individual residential architecture, a body boundary which relates the space between outside and inside, between the external and the self, between “you” and “I”; at the same time a specific and abstract filter; an interface where the dress is the container and the body its content.

Parallel Computer Vision Algorithms for Graphics Processing Units

La evolución de los teléfonos móviles inteligentes, dotados de cámaras digitales, está provocando una creciente demanda de aplicaciones cada vez más complejas que necesitan algoritmos de visión artificial en tiempo real; puesto que el tamaño de las señales de vídeo no hace sino aumentar y en cambio el rendimiento de los procesadores de un solo núcleo se ha estancado, los nuevos algoritmos que se diseñen para visión artificial han de ser paralelos para poder ejecutarse en múltiples procesadores y ser computacionalmente escalables. Una de las clases de procesadores más interesantes en la actualidad se encuentra en las tarjetas gráficas (GPU), que son dispositivos que ofrecen un alto grado de paralelismo, un excelente rendimiento numérico y una creciente versatilidad, lo que los hace interesantes para llevar a cabo computación científica. En esta tesis se exploran dos aplicaciones de visión artificial que revisten una gran complejidad computacional y no pueden ser ejecutadas en tiempo real empleando procesadores tradicionales. En cambio, como se demuestra en esta tesis, la paralelización de las distintas subtareas y su implementación sobre una GPU arrojan los resultados deseados de ejecución con tasas de refresco interactivas. Asimismo, se propone una técnica para la evaluación rápida de funciones de complejidad arbitraria especialmente indicada para su uso en una GPU. En primer lugar se estudia la aplicación de técnicas de síntesis de imágenes virtuales a partir de únicamente dos cámaras lejanas y no paralelas—en contraste con la configuración habitual en TV 3D de cámaras cercanas y paralelas—con información de color y profundidad. Empleando filtros de mediana modificados para la elaboración de un mapa de profundidad virtual y proyecciones inversas, se comprueba que estas técnicas son adecuadas para una libre elección del punto de vista. Además, se demuestra que la codificación de la información de profundidad con respecto a un sistema de referencia global es sumamente perjudicial y debería ser evitada. Por otro lado se propone un sistema de detección de objetos móviles basado en técnicas de estimación de densidad con funciones locales. Este tipo de técnicas es muy adecuada para el modelado de escenas complejas con fondos multimodales, pero ha recibido poco uso debido a su gran complejidad computacional. El sistema propuesto, implementado en tiempo real sobre una GPU, incluye propuestas para la estimación dinámica de los anchos de banda de las funciones locales, actualización selectiva del modelo de fondo, actualización de la posición de las muestras de referencia del modelo de primer plano empleando un filtro de partículas multirregión y selección automática de regiones de interés para reducir el coste computacional. Los resultados, evaluados sobre diversas bases de datos y comparados con otros algoritmos del estado del arte, demuestran la gran versatilidad y calidad de la propuesta. Finalmente se propone un método para la aproximación de funciones arbitrarias empleando funciones continuas lineales a tramos, especialmente indicada para su implementación en una GPU mediante el uso de las unidades de filtraje de texturas, normalmente no utilizadas para cómputo numérico. La propuesta incluye un riguroso análisis matemático del error cometido en la aproximación en función del número de muestras empleadas, así como un método para la obtención de una partición cuasióptima del dominio de la función para minimizar el error. ABSTRACT The evolution of smartphones, all equipped with digital cameras, is driving a growing demand for ever more complex applications that need to rely on real-time computer vision algorithms. However, video signals are only increasing in size, whereas the performance of single-core processors has somewhat stagnated in the past few years. Consequently, new computer vision algorithms will need to be parallel to run on multiple processors and be computationally scalable. One of the most promising classes of processors nowadays can be found in graphics processing units (GPU). These are devices offering a high parallelism degree, excellent numerical performance and increasing versatility, which makes them interesting to run scientific computations. In this thesis, we explore two computer vision applications with a high computational complexity that precludes them from running in real time on traditional uniprocessors. However, we show that by parallelizing subtasks and implementing them on a GPU, both applications attain their goals of running at interactive frame rates. In addition, we propose a technique for fast evaluation of arbitrarily complex functions, specially designed for GPU implementation. First, we explore the application of depth-image–based rendering techniques to the unusual configuration of two convergent, wide baseline cameras, in contrast to the usual configuration used in 3D TV, which are narrow baseline, parallel cameras. By using a backward mapping approach with a depth inpainting scheme based on median filters, we show that these techniques are adequate for free viewpoint video applications. In addition, we show that referring depth information to a global reference system is ill-advised and should be avoided. Then, we propose a background subtraction system based on kernel density estimation techniques. These techniques are very adequate for modelling complex scenes featuring multimodal backgrounds, but have not been so popular due to their huge computational and memory complexity. The proposed system, implemented in real time on a GPU, features novel proposals for dynamic kernel bandwidth estimation for the background model, selective update of the background model, update of the position of reference samples of the foreground model using a multi-region particle filter, and automatic selection of regions of interest to reduce computational cost. The results, evaluated on several databases and compared to other state-of-the-art algorithms, demonstrate the high quality and versatility of our proposal. Finally, we propose a general method for the approximation of arbitrarily complex functions using continuous piecewise linear functions, specially formulated for GPU implementation by leveraging their texture filtering units, normally unused for numerical computation. Our proposal features a rigorous mathematical analysis of the approximation error in function of the number of samples, as well as a method to obtain a suboptimal partition of the domain of the function to minimize approximation error.