A Comparison of Hardware Implementations for Low-Level Vision Algorithms

Early and intermediate vision algorithms, such as smoothing and discontinuity detection, are often implemented on general-purpose serial, and more recently, parallel computers. Special-purpose hardware implementations of low-level vision algorithms may be needed to achieve real-time processing. This memo reviews and analyzes some hardware implementations of low-level vision algorithms. Two types of hardware implementations are considered: the digital signal processing chips of Ruetz (and Broderson) and the analog VLSI circuits of Carver Mead. The advantages and disadvantages of these two approaches for producing a general, real-time vision system are considered.

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.

3D Visual Data-Driven Spatiotemporal Deformations for Non-Rigid Object Grasping Using Robot Hands

Sensing techniques are important for solving problems of uncertainty inherent to intelligent grasping tasks. The main goal here is to present a visual sensing system based on range imaging technology for robot manipulation of non-rigid objects. Our proposal provides a suitable visual perception system of complex grasping tasks to support a robot controller when other sensor systems, such as tactile and force, are not able to obtain useful data relevant to the grasping manipulation task. In particular, a new visual approach based on RGBD data was implemented to help a robot controller carry out intelligent manipulation tasks with flexible objects. The proposed method supervises the interaction between the grasped object and the robot hand in order to avoid poor contact between the fingertips and an object when there is neither force nor pressure data. This new approach is also used to measure changes to the shape of an object’s surfaces and so allows us to find deformations caused by inappropriate pressure being applied by the hand’s fingers. Test was carried out for grasping tasks involving several flexible household objects with a multi-fingered robot hand working in real time. Our approach generates pulses from the deformation detection method and sends an event message to the robot controller when surface deformation is detected. In comparison with other methods, the obtained results reveal that our visual pipeline does not use deformations models of objects and materials, as well as the approach works well both planar and 3D household objects in real time. In addition, our method does not depend on the pose of the robot hand because the location of the reference system is computed from a recognition process of a pattern located place at the robot forearm. The presented experiments demonstrate that the proposed method accomplishes a good monitoring of grasping task with several objects and different grasping configurations in indoor environments.

The Machine Vision Toolbox : A MATLAB Toolbox for vision and vision-based control

The article described an open-source toolbox for machine vision called Machine Vision Toolbox (MVT). MVT includes more than 60 functions including image file reading and writing, acquisition, display, filtering, blob, point and line feature extraction, mathematical morphology, homographies, visual Jacobians, camera calibration, and color space conversion. MVT can be used for research into machine vision but is also versatile enough to be usable for real-time work and even control. MVT, combined with MATLAB and a model workstation computer, is a useful and convenient environment for the investigation of machine vision algorithms. The article illustrated the use of a subset of toolbox functions for some typical problems and described MVT operations including the simulation of a complete image-based visual servo system.

Machine vision toolbox for MATLAB, release 3 [Software]

The Toolbox, combined with MATLAB ® and a modern workstation computer, is a useful and convenient environment for investigation of machine vision algorithms. For modest image sizes the processing rate can be sufficiently ``real-time'' to allow for closed-loop control. Focus of attention methods such as dynamic windowing (not provided) can be used to increase the processing rate. With input from a firewire or web camera (support provided) and output to a robot (not provided) it would be possible to implement a visual servo system entirely in MATLAB. Provides many functions that are useful in machine vision and vision-based control. Useful for photometry, photogrammetry, colorimetry. It includes over 100 functions spanning operations such as image file reading and writing, acquisition, display, filtering, blob, point and line feature extraction, mathematical morphology, homographies, visual Jacobians, camera calibration and color space conversion.

An SIMD machine for low-level vision

This paper presents an SIMD machine which has been tuned to execute low-level vision algorithms employing the relaxation labeling paradigm. Novel features of the design include: 1. (1) a communication scheme capable of window accessing under a single instruction. 2. (2) flexible I/O instructions to load overlapped data segments; and 3. (3) data-conditional instructions which can be nested to an arbitrary degree. A time analysis of the stereo correspondence problem, as implemented on a simulated version of the machine using the probabilistic relaxation technique, shows a speed up of almost N2 for an N × N array of PEs.

Dissecting neural circuits for vision in nonhuman primates using fMRI-guided electrophysiology and optogenetics

The visual system is a remarkable platform that evolved to solve difficult computational problems such as detection, recognition, and classification of objects. Of great interest is the face-processing network, a sub-system buried deep in the temporal lobe, dedicated for analyzing specific type of objects (faces). In this thesis, I focus on the problem of face detection by the face-processing network. Insights obtained from years of developing computer-vision algorithms to solve this task have suggested that it may be efficiently and effectively solved by detection and integration of local contrast features. Does the brain use a similar strategy? To answer this question, I embark on a journey that takes me through the development and optimization of dedicated tools for targeting and perturbing deep brain structures. Data collected using MR-guided electrophysiology in early face-processing regions was found to have strong selectivity for contrast features, similar to ones used by artificial systems. While individual cells were tuned for only a small subset of features, the population as a whole encoded the full spectrum of features that are predictive to the presence of a face in an image. Together with additional evidence, my results suggest a possible computational mechanism for face detection in early face processing regions. To move from correlation to causation, I focus on adopting an emergent technology for perturbing brain activity using light: optogenetics. While this technique has the potential to overcome problems associated with the de-facto way of brain stimulation (electrical microstimulation), many open questions remain about its applicability and effectiveness for perturbing the non-human primate (NHP) brain. In a set of experiments, I use viral vectors to deliver genetically encoded optogenetic constructs to the frontal eye field and faceselective regions in NHP and examine their effects side-by-side with electrical microstimulation to assess their effectiveness in perturbing neural activity as well as behavior. Results suggest that cells are robustly and strongly modulated upon light delivery and that such perturbation can modulate and even initiate motor behavior, thus, paving the way for future explorations that may apply these tools to study connectivity and information flow in the face processing network.

The smart vision local navigation aid for blind and visually impaired persons

The SmartVision prototype is a small, cheap and easily wearable navigation aid for blind and visually impaired persons. Its functionality addresses global navigation for guiding the user to some destiny, and local navigation for negotiating paths, sidewalks and corridors, with avoidance of static as well as moving obstacles. Local navigation applies to both in- and outdoor situations. In this article we focus on local navigation: the detection of path borders and obstacles in front of the user and just beyond the reach of the white cane, such that the user can be assisted in centering on the path and alerted to looming hazards. Using a stereo camera worn at chest height, a portable computer in a shoulder-strapped pouch or pocket and only one earphone or small speaker, the system is inconspicuous, it is no hindrence while walking with the cane, and it does not block normal surround sounds. The vision algorithms are optimised such that the system can work at a few frames per second.

Quantum of Vision

Visual inputs to artificial and biological visual systems are often quantized: cameras accumulate photons from the visual world, and the brain receives action potentials from visual sensory neurons. Collecting more information quanta leads to a longer acquisition time and better performance. In many visual tasks, collecting a small number of quanta is sufficient to solve the task well. The ability to determine the right number of quanta is pivotal in situations where visual information is costly to obtain, such as photon-starved or time-critical environments. In these situations, conventional vision systems that always collect a fixed and large amount of information are infeasible. I develop a framework that judiciously determines the number of information quanta to observe based on the cost of observation and the requirement for accuracy. The framework implements the optimal speed versus accuracy tradeoff when two assumptions are met, namely that the task is fully specified probabilistically and constant over time. I also extend the framework to address scenarios that violate the assumptions. I deploy the framework to three recognition tasks: visual search (where both assumptions are satisfied), scotopic visual recognition (where the model is not specified), and visual discrimination with unknown stimulus onset (where the model is dynamic over time). Scotopic classification experiments suggest that the framework leads to dramatic improvement in photon-efficiency compared to conventional computer vision algorithms. Human psychophysics experiments confirmed that the framework provides a parsimonious and versatile explanation for human behavior under time pressure in both static and dynamic environments.

An accurate and reliable stereo matching algorithm incorporating the rank constraint

A fundamental problem faced by stereo vision algorithms is that of determining correspondences between two images which comprise a stereo pair. This paper presents work towards the development of a new matching algorithm, based on the rank transform. This algorithm makes use of both area-based and edge-based information, and is therefore referred to as a hybrid algorithm. In addition, this algorithm uses a number of matching constraints,including the novel rank constraint. Results obtained using a number of test pairs show that the matching algorithm is capable of removing a significant proportion of invalid matches. The accuracy of matching in the vicinity of edges is also improved.

A hybrid stereo matching algorithm incorporating the rank constraint

A fundamental problem faced by stereo vision algorithms is that of determining correspondences between two images which comprise a stereo pair. This paper presents work towards the development of a new matching algorithm, based on the rank transform. This algorithm makes use of both area-based and edge-based information, and is therefore referred to as a hybrid algorithm. In addition, this algorithm uses a number of matching constraints, including the novel rank constraint. Results obtained using a number of test pairs show that the matching algorithm is capable of removing most invalid matches. The accuracy of matching in the vicinity of edges is also improved.

Simultaneous underwater visibility assessment, enhancement and improved stereo

Vision-based underwater navigation and obstacle avoidance demands robust computer vision algorithms, particularly for operation in turbid water with reduced visibility. This paper describes a novel method for the simultaneous underwater image quality assessment, visibility enhancement and disparity computation to increase stereo range resolution under dynamic, natural lighting and turbid conditions. The technique estimates the visibility properties from a sparse 3D map of the original degraded image using a physical underwater light attenuation model. Firstly, an iterated distance-adaptive image contrast enhancement enables a dense disparity computation and visibility estimation. Secondly, using a light attenuation model for ocean water, a color corrected stereo underwater image is obtained along with a visibility distance estimate. Experimental results in shallow, naturally lit, high-turbidity coastal environments show the proposed technique improves range estimation over the original images as well as image quality and color for habitat classification. Furthermore, the recursiveness and robustness of the technique allows implementation onboard an Autonomous Underwater Vehicle for improving navigation and obstacle avoidance performance.

Searching for people using semantic soft biometric descriptions

It is not uncommon to hear a person of interest described by their height, build, and clothing (i.e. type and colour). These semantic descriptions are commonly used by people to describe others, as they are quick to communicate and easy to understand. However such queries are not easily utilised within intelligent video surveillance systems, as they are difficult to transform into a representation that can be utilised by computer vision algorithms. In this paper we propose a novel approach that transforms such a semantic query into an avatar in the form of a channel representation that is searchable within a video stream. We show how spatial, colour and prior information (person shape) can be incorporated into the channel representation to locate a target using a particle-filter like approach. We demonstrate state-of-the-art performance for locating a subject in video based on a description, achieving a relative performance improvement of 46.7% over the baseline. We also apply this approach to person re-detection, and show that the approach can be used to re-detect a person in a video steam without the use of person detection.

Linear Chromatic Adaptation Transform Based on Delaunay Triangulation

Computer vision algorithms that use color information require color constant images to operate correctly. Color constancy of the images is usually achieved in two steps: first the illuminant is detected and then image is transformed with the chromatic adaptation transform ( CAT). Existing CAT methods use a single transformation matrix for all the colors of the input image. The method proposed in this paper requires multiple corresponding color pairs between source and target illuminants given by patches of the Macbeth color checker. It uses Delaunay triangulation to divide the color gamut of the input image into small triangles. Each color of the input image is associated with the triangle containing the color point and transformed with a full linear model associated with the triangle. Full linear model is used because diagonal models are known to be inaccurate if channel color matching functions do not have narrow peaks. Objective evaluation showed that the proposed method outperforms existing CAT methods by more than 21%; that is, it performs statistically significantly better than other existing methods.