Scalable software architecture for on-line multi-camera video processing

In this paper we present a scalable software architecture for on-line multi-camera video processing, that guarantees a good trade off between computational power, scalability and flexibility. The software system is modular and its main blocks are the Processing Units (PUs), and the Central Unit. The Central Unit works as a supervisor of the running PUs and each PU manages the acquisition phase and the processing phase. Furthermore, an approach to easily parallelize the desired processing application has been presented. In this paper, as case study, we apply the proposed software architecture to a multi-camera system in order to efficiently manage multiple 2D object detection modules in a real-time scenario. System performance has been evaluated under different load conditions such as number of cameras and image sizes. The results show that the software architecture scales well with the number of camera and can easily works with different image formats respecting the real time constraints. Moreover, the parallelization approach can be used in order to speed up the processing tasks with a low level of overhead

3D video quality of experience - influence of scale and crosstalk

This paper gives an overview of three recent studies by the authors on the topic of 3D video Quality of Experience (QoE). Two of studies [1,2] investigated different psychological dimension that may be needed for describing 3D video QoE and the third the visibility and annoyance of crosstalk[3]. The results shows that the video quality scale could be sufficient for evaluating S3D video experience for coding and spatial resolution reduction distortions. It was also confirmed that with a more complex mixture of degradations more than one scale should be used to capture the QoE in these cases. The study found a linear relationship between the perceived crosstalk and the amount of crosstalk.

Analysis of quality of experience in 3D video systems

Esta tesis presenta un estudio exhaustivo sobre la evaluación de la calidad de experiencia (QoE, del inglés Quality of Experience) percibida por los usuarios de sistemas de vídeo 3D, analizando el impacto de los efectos introducidos por todos los elementos de la cadena de procesamiento de vídeo 3D. Por lo tanto, se presentan varias pruebas de evaluación subjetiva específicamente diseñadas para evaluar los sistemas considerados, teniendo en cuenta todos los factores perceptuales relacionados con la experiencia visual tridimensional, tales como la percepción de profundidad y la molestia visual. Concretamente, se describe un test subjetivo basado en la evaluación de degradaciones típicas que pueden aparecer en el proceso de creación de contenidos de vídeo 3D, por ejemplo debidas a calibraciones incorrectas de las cámaras o a algoritmos de procesamiento de la señal de vídeo (p. ej., conversión de 2D a 3D). Además, se presenta el proceso de generación de una base de datos de vídeos estereoscópicos de alta calidad, disponible gratuitamente para la comunidad investigadora y que ha sido utilizada ampliamente en diferentes trabajos relacionados con vídeo 3D. Asimismo, se presenta otro estudio subjetivo, realizado entre varios laboratorios, con el que se analiza el impacto de degradaciones causadas por la codificación de vídeo, así como diversos formatos de representación de vídeo 3D. Igualmente, se describen tres pruebas subjetivas centradas en el estudio de posibles efectos causados por la transmisión de vídeo 3D a través de redes de televisión sobre IP (IPTV, del inglés Internet Protocol Television) y de sistemas de streaming adaptativo de vídeo. Para estos casos, se ha propuesto una innovadora metodología de evaluación subjetiva de calidad vídeo, denominada Content-Immersive Evaluation of Transmission Impairments (CIETI), diseñada específicamente para evaluar eventos de transmisión simulando condiciones realistas de visualización de vídeo en ámbitos domésticos, con el fin de obtener conclusiones más representativas sobre la experiencia visual de los usuarios finales. Finalmente, se exponen dos experimentos subjetivos comparando varias tecnologías actuales de televisores 3D disponibles en el mercado de consumo y evaluando factores perceptuales de sistemas Super Multiview Video (SMV), previstos a ser la tecnología futura de televisores 3D de consumo, gracias a una prometedora visualización de contenido 3D sin necesidad de gafas específicas. El trabajo presentado en esta tesis ha permitido entender los factores perceptuales y técnicos relacionados con el procesamiento y visualización de contenidos de vídeo 3D, que pueden ser de utilidad en el desarrollo de nuevas tecnologías y técnicas de evaluación de la QoE, tanto metodologías subjetivas como métricas objetivas. ABSTRACT This thesis presents a comprehensive study of the evaluation of the Quality of Experience (QoE) perceived by the users of 3D video systems, analyzing the impact of effects introduced by all the elements of the 3D video processing chain. Therefore, various subjective assessment tests are presented, particularly designed to evaluate the systems under consideration, and taking into account all the perceptual factors related to the 3D visual experience, such as depth perception and visual discomfort. In particular, a subjective test is presented, based on evaluating typical degradations that may appear during the content creation, for instance due to incorrect camera calibration or video processing algorithms (e.g., 2D to 3D conversion). Moreover, the process of generation of a high-quality dataset of 3D stereoscopic videos is described, which is freely available for the research community, and has been already widely used in different works related with 3D video. In addition, another inter-laboratory subjective study is presented analyzing the impact of coding impairments and representation formats of stereoscopic video. Also, three subjective tests are presented studying the effects of transmission events that take place in Internet Protocol Television (IPTV) networks and adaptive streaming scenarios for 3D video. For these cases, a novel subjective evaluation methodology, called Content-Immersive Evaluation of Transmission Impairments (CIETI), was proposed, which was especially designed to evaluate transmission events simulating realistic home-viewing conditions, to obtain more representative conclusions about the visual experience of the end users. Finally, two subjective experiments are exposed comparing various current 3D displays available in the consumer market, and evaluating perceptual factors of Super Multiview Video (SMV) systems, expected to be the future technology for consumer 3D displays thanks to a promising visualization of 3D content without specific glasses. The work presented in this thesis has allowed to understand perceptual and technical factors related to the processing and visualization of 3D video content, which may be useful in the development of new technologies and approaches for QoE evaluation, both subjective methodologies and objective metrics.

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.

An approach to hand biometrics in mobile devices

This paper focuses on hand biometrics applied to images acquired from a mobile device. The system offers the possibility of identifying individuals based on features extracted from hand pictures obtained with a low-quality camera embedded on a mobile device. Furthermore, the acquisitions have been carried out regardless illumination control, orientation, distance to camera, and similar aspects. In addition, the whole system has been tested with an owned database. Finally, the results obtained (6.0% ± 0.2) and the algorithm structure are both promising in relation to a posterior mobile implementation

Applications of image-based, multi-view stereo reconstruction methods = (Aplicaciones de los métodos de reconstrucción estéreo multivista)

These slides present several 3-D reconstruction methods to obtain the geometric structure of a scene that is viewed by multiple cameras. We focus on the combination of the geometric modeling in the image formation process with the use of standard optimization tools to estimate the characteristic parameters that describe the geometry of the 3-D scene. In particular, linear, non-linear and robust methods to estimate the monocular and epipolar geometry are introduced as cornerstones to generate 3-D reconstructions with multiple cameras. Some examples of systems that use this constructive strategy are Bundler, PhotoSynth, VideoSurfing, etc., which are able to obtain 3-D reconstructions with several hundreds or thousands of cameras. En esta presentación se tratan varios métodos de reconstrucción 3-D para la obtención de la estructura geométrica de una escena que es visualizada por varias cámaras. Se enfatiza la combinación de modelado geométrico del proceso de formación de la imagen con el uso de herramientas estándar de optimización para estimar los parámetros característicos que describen la geometría de la escena 3-D. En concreto, se presentan métodos de estimación lineales, no lineales y robustos de las geometrías monocular y epipolar como punto de partida para generar reconstrucciones con tres o más cámaras. Algunos ejemplos de sistemas que utilizan este enfoque constructivo son Bundler, PhotoSynth, VideoSurfing, etc., los cuales, en la práctica pueden llegar a reconstruir una escena con varios cientos o miles de cámaras.

Architecture for Text Sign Localization and Recognition

This paper describes a low complexity strategy for detecting and recognizing text signs automatically. Traditional approaches use large image algorithms for detecting the text sign, followed by the application of an Optical Character Recognition (OCR) algorithm in the previously identified areas. This paper proposes a new architecture that applies the OCR to a whole lightly treated image and then carries out the text detection process of the OCR output. The strategy presented in this paper significantly reduces the processing time required for text localization in an image, while guaranteeing a high recognition rate. This strategy will facilitate the incorporation of video processing-based applications into the automatic detection of text sign similar to that of a smartphone. These applications will increase the autonomy of visually impaired people in their daily life.

Modelado en alto nivel con SystemC

El presente proyecto final de carrera titulado “Modelado de alto nivel con SystemC” tiene como objetivo principal el modelado de algunos módulos de un codificador de vídeo MPEG-2 utilizando el lenguaje de descripción de sistemas igitales SystemC con un nivel de abstracción TLM o Transaction Level Modeling. SystemC es un lenguaje de descripción de sistemas digitales basado en C++. En él hay un conjunto de rutinas y librerías que implementan tipos de datos, estructuras y procesos especiales para el modelado de sistemas digitales. Su descripción se puede consultar en [GLMS02] El nivel de abstracción TLM se caracteriza por separar la comunicación entre los módulos de su funcionalidad. Este nivel de abstracción hace un mayor énfasis en la funcionalidad de la comunicación entre los módulos (de donde a donde van datos) que la implementación exacta de la misma. En los documentos [RSPF] y [HG] se describen el TLM y un ejemplo de implementación. La arquitectura del modelo se basa en el codificador MVIP-2 descrito en [Gar04], de dicho modelo, los módulos implementados son: · IVIDEOH: módulo que realiza un filtrado del vídeo de entrada en la dimensión horizontal y guarda en memoria el video filtrado. · IVIDEOV: módulo que lee de la memoria el vídeo filtrado por IVIDEOH, realiza el filtrado en la dimensión horizontal y escribe el video filtrado en memoria. · DCT: módulo que lee el video filtrado por IVIDEOV, hace la transformada discreta del coseno y guarda el vídeo transformado en la memoria. · QUANT: módulo que lee el video transformado por DCT, lo cuantifica y guarda el resultado en la memoria. · IQUANT: módulo que lee el video cuantificado por QUANT, realiza la cuantificación inversa y guarda el resultado en memoria. · IDCT: módulo que lee el video procesado por IQUANT, realiza la transformada inversa del coseno y guarda el resultado en memoria. · IMEM: módulo que hace de interfaz entre los módulos anteriores y la memoria. Gestiona las peticiones simultáneas de acceso a la memoria y asegura el acceso exclusivo a la memoria en cada instante de tiempo. Todos estos módulos aparecen en gris en la siguiente figura en la que se muestra la arquitectura del modelo: Figura 1. Arquitectura del modelo (VER PDF DEL PFC) En figura también aparecen unos módulos en blanco, dichos módulos son de pruebas y se han añadido para realizar simulaciones y probar los módulos del modelo: · CAMARA: módulo que simula una cámara en blanco y negro, lee la luminancia de un fichero de vídeo y lo envía al modelo a través de una FIFO. · FIFO: hace de interfaz entre la cámara y el modelo, guarda los datos que envía la cámara hasta que IVIDEOH los lee. · CONTROL: módulo que se encarga de controlar los módulos que procesan el vídeo, estos le indican cuando terminan de procesar un frame de vídeo y este módulo se encarga de iniciar los módulos que sean necesarios para seguir con la codificación. Este módulo se encarga del correcto secuenciamiento de los módulos procesadores de vídeo. · RAM: módulo que simula una memoria RAM, incluye un retardo programable en el acceso. Para las pruebas también se han generado ficheros de vídeo con el resultado de cada módulo procesador de vídeo, ficheros con mensajes y un fichero de trazas en el que se muestra el secuenciamiento de los procesadores. Como resultado del trabajo en el presente PFC se puede concluir que SystemC permite el modelado de sistemas digitales con bastante sencillez (hace falta conocimientos previos de C++ y programación orientada objetos) y permite la realización de modelos con un nivel de abstracción mayor a RTL, el habitual en Verilog y VHDL, en el caso del presente PFC, el TLM. ABSTRACT This final career project titled “High level modeling with SystemC” have as main objective the modeling of some of the modules of an MPEG-2 video coder using the SystemC digital systems description language at the TLM or Transaction Level Modeling abstraction level. SystemC is a digital systems description language based in C++. It contains routines and libraries that define special data types, structures and process to model digital systems. There is a complete description of the SystemC language in the document [GLMS02]. The main characteristic of TLM abstraction level is that it separates the communication among modules of their functionality. This abstraction level puts a higher emphasis in the functionality of the communication (from where to where the data go) than the exact implementation of it. The TLM and an example are described in the documents [RSPF] and [HG]. The architecture of the model is based in the MVIP-2 video coder (described in the document [Gar04]) The modeled modules are: · IVIDEOH: module that filter the video input in the horizontal dimension. It saves the filtered video in the memory. · IVIDEOV: module that read the IVIDEOH filtered video, filter it in the vertical dimension and save the filtered video in the memory. · DCT: module that read the IVIDEOV filtered video, do the discrete cosine transform and save the transformed video in the memory. · QUANT: module that read the DCT transformed video, quantify it and save the quantified video in the memory. · IQUANT: module that read the QUANT processed video, do the inverse quantification and save the result in the memory. · IDCT: module that read the IQUANT processed video, do the inverse cosine transform and save the result in the memory. · IMEM: this module is the interface between the modules described previously and the memory. It manage the simultaneous accesses to the memory and ensure an unique access at each instant of time All this modules are included in grey in the following figure (SEE PDF OF PFC). This figure shows the architecture of the model: Figure 1. Architecture of the model This figure also includes other modules in white, these modules have been added to the model in order to simulate and prove the modules of the model: · CAMARA: simulates a black and white video camera, it reads the luminance of a video file and sends it to the model through a FIFO. · FIFO: is the interface between the camera and the model, it saves the video data sent by the camera until the IVIDEOH module reads it. · CONTROL: controls the modules that process the video. These modules indicate the CONTROL module when they have finished the processing of a video frame. The CONTROL module, then, init the necessary modules to continue with the video coding. This module is responsible of the right sequence of the video processing modules. · RAM: it simulates a RAM memory; it also simulates a programmable delay in the access to the memory. It has been generated video files, text files and a trace file to check the correct function of the model. The trace file shows the sequence of the video processing modules. As a result of the present final career project, it can be deduced that it is quite easy to model digital systems with SystemC (it is only needed previous knowledge of C++ and object oriented programming) and it also allow the modeling with a level of abstraction higher than the RTL used in Verilog and VHDL, in the case of the present final career project, the TLM.

Multi-resolution model-based traffic sign detection and tracking

In this paper we propose an innovative approach to tackle the problem of traffic sign detection using a computer vision algorithm and taking into account real-time operation constraints, trying to establish intelligent strategies to simplify as much as possible the algorithm complexity and to speed up the process. Firstly, a set of candidates is generated according to a color segmentation stage, followed by a region analysis strategy, where spatial characteristic of previously detected objects are taken into account. Finally, temporal coherence is introduced by means of a tracking scheme, performed using a Kalman filter for each potential candidate. Taking into consideration time constraints, efficiency is achieved two-fold: on the one side, a multi-resolution strategy is adopted for segmentation, where global operation will be applied only to low-resolution images, increasing the resolution to the maximum only when a potential road sign is being tracked. On the other side, we take advantage of the expected spacing between traffic signs. Namely, the tracking of objects of interest allows to generate inhibition areas, which are those ones where no new traffic signs are expected to appear due to the existence of a TS in the neighborhood. The proposed solution has been tested with real sequences in both urban areas and highways, and proved to achieve higher computational efficiency, especially as a result of the multi-resolution approach.

Ampliación y mejora de aplicativo de análisis forense de imágenes

Ampliación de software dedicado al análisis de imágenes mediante la introducción de nuevas opciones en el procesamiento de video digital, mejoras en la interacción con el usuario. Para ello se ha estudiado el funcionamiento de la aplicación, integrando el lenguaje Python como herramienta de gestión y ejecución de la aplicación. En esta parte de la aplicación se ha integrado: - Traducción de la UI a una versión castellana. - Modificación y eliminación de cualquier filtro añadido para el procesamiento de video, no únicamente el último. - Descripciones de puntero y en la barra de estado de elementos de la aplicación. - Iconos en la barra de herramientas de los filtros añadidos más importantes. Por la otra parte, la del tratamiento digital de video, Avisynth se dispone como el eje de estudio, el cuál ejecuta sobre lenguaje de bajo nivel (C++) las operaciones pertinentes a través de librerías de enlace dinámico o *.dll. Las nuevas funcionalidades son: Convolución matricial, filtro de media adaptativa, DCT, ajustes de video generales, en formato RGB o YUV, rotaciones, cambios de perspectiva y filtrado en frecuencia. ABSTRACT. Improvement about a digital image processing software, creating new options in digital video processing or the user interaction. For this porpuse, we have integrated the application language,Python, as the tool to the application management and execution. In this part of the application has been integrated: - Translation of the UI: Spanish version. - Modifying and removing any added filter for video processing, not just the last. - Descriptions for the pointer and the status bar of the application. - New icons on the toolbar of the most important filters added. On the other hand, Avisynth was used tool for the digital video processing, which runs on low-level language (C ++) for a quickly and to improve the video operations. The new introduced filters are: Matrix Convolution, adaptive median filter, DCT, general video settings on RGB or YUV format, rotations, changes in perspective and frequency filtering.

Video-based tasks for emotional processing rehabilitation in schizophrenia

Schizophrenia is a mental disorder characterized by a breakdown of cognitive processes and by a deficit of typi-cal emotional responses. Effectiveness of computerized task has been demonstrated in the field of cognitive rehabilitation. However, current rehabilitation programs based on virtual environments normally focus on higher cognitive functions, not covering social cognition training. This paper presents a set of video-based tasks specifically designed for the rehabilita-tion of emotional processing deficits in patients in early stages of schizophrenia or schizoaffective disorders. These tasks are part of the Mental Health program of Guttmann NeuroPer-sonalTrainer® cognitive tele-rehabilitation platform, and entail innovation both from a clinical and technological per-spective in relation with former traditional therapeutic con-tents.

Video analysis based vehicle detection and tracking using an MCMC sampling framework

This article presents a probabilistic method for vehicle detection and tracking through the analysis of monocular images obtained from a vehicle-mounted camera. The method is designed to address the main shortcomings of traditional particle filtering approaches, namely Bayesian methods based on importance sampling, for use in traffic environments. These methods do not scale well when the dimensionality of the feature space grows, which creates significant limitations when tracking multiple objects. Alternatively, the proposed method is based on a Markov chain Monte Carlo (MCMC) approach, which allows efficient sampling of the feature space. The method involves important contributions in both the motion and the observation models of the tracker. Indeed, as opposed to particle filter-based tracking methods in the literature, which typically resort to observation models based on appearance or template matching, in this study a likelihood model that combines appearance analysis with information from motion parallax is introduced. Regarding the motion model, a new interaction treatment is defined based on Markov random fields (MRF) that allows for the handling of possible inter-dependencies in vehicle trajectories. As for vehicle detection, the method relies on a supervised classification stage using support vector machines (SVM). The contribution in this field is twofold. First, a new descriptor based on the analysis of gradient orientations in concentric rectangles is dened. This descriptor involves a much smaller feature space compared to traditional descriptors, which are too costly for real-time applications. Second, a new vehicle image database is generated to train the SVM and made public. The proposed vehicle detection and tracking method is proven to outperform existing methods and to successfully handle challenging situations in the test sequences.

A set of canonical problems in sloshing. Part 0: Experimental setup and data processing

In a series of attempts to research and document relevant sloshing type phenomena, a series of experiments have been conducted. The aim of this paper is to describe the setup and data processing of such experiments. A sloshing tank is subjected to angular motion. As a result pressure registers are obtained at several locations, together with the motion data, torque and a collection of image and video information. The experimental rig and the data acquisition systems are described. Useful information for experimental sloshing research practitioners is provided. This information is related to the liquids used in the experiments, the dying techniques, tank building processes, synchronization of acquisition systems, etc. A new procedure for reconstructing experimental data, that takes into account experimental uncertainties, is presented. This procedure is based on a least squares spline approximation of the data. Based on a deterministic approach to the first sloshing wave impact event in a sloshing experiment, an uncertainty analysis procedure of the associated first pressure peak value is described.

A graph-based approach for latency modeling and optimization in multiview video encoding

We present a novel framework for encoding latency analysis of arbitrary multiview video coding prediction structures. This framework avoids the need to consider an specific encoder architecture for encoding latency analysis by assuming an unlimited processing capacity on the multiview encoder. Under this assumption, only the influence of the prediction structure and the processing times have to be considered, and the encoding latency is solved systematically by means of a graph model. The results obtained with this model are valid for a multiview encoder with sufficient processing capacity and serve as a lower bound otherwise. Furthermore, with the objective of low latency encoder design with low penalty on rate-distortion performance, the graph model allows us to identify the prediction relationships that add higher encoding latency to the encoder. Experimental results for JMVM prediction structures illustrate how low latency prediction structures with a low rate-distortion penalty can be derived in a systematic manner using the new model.

Area-delay Trade-offs of Texture Decompressors for a Graphics Processing Unit

Graphics Processing Units have become a booster for the microelectronics industry. However, due to intellectual property issues, there is a serious lack of information on implementation details of the hardware architecture that is behind GPUs. For instance, the way texture is handled and decompressed in a GPU to reduce bandwidth usage has never been dealt with in depth from a hardware point of view. This work addresses a comparative study on the hardware implementation of different texture decompression algorithms for both conventional (PCs and video game consoles) and mobile platforms. Circuit synthesis is performed targeting both a reconfigurable hardware platform and a 90nm standard cell library. Area-delay trade-offs have been extensively analyzed, which allows us to compare the complexity of decompressors and thus determine suitability of algorithms for systems with limited hardware resources.