A significant amount of research in the field of stereo vision has been published in the past decade. Considerable progress has been made in improving accuracy of results as well as achieving real-time performance in obtaining those results. Although much of the literature does not address it, many applications are sensitive to the tradeoff between accuracy and speed that exists among stereo vision algorithms. Overall, this work aims to organize existing efforts and encourage new ones in the development of stereo vision algorithms for resource limited systems. It does this through a review of the status quo as well as providing both software and hardware designs of new stereo vision algorithms that offer an efficient tradeoff between speed and accuracy. A comprehensive review and analysis of stereo vision algorithms is provided with specific emphasis on real-time performance and suitability for resource limited systems. An attempt has been made to compile and present accuracy and runtime performance data for all stereo vision algorithms developed in the past decade. The tradeoff in accuracy that is typically made to achieve real-time performance is examined with an example of an existing highly accurate stereo vision that is modified to see how much speedup can be achieved. Two new stereo vision algorithms, GA Spline and Profile Shape Matching, are presented with a hardware design of the latter also being provided, making Profile Shape Matching available to both embedded processor-based and programmable hardware-based resource limited systems.

Suffisante pour exécuter les algorithmes à la cadence de ces capteurs d’images performants, tout en gardant une faible consommation d’énergie. Les systèmes monoprocesseur n’arrivent plus à satisfaire les exigences de ce domaine. Ainsi, grâce aux avancées technologiques et en s’appuyant sur de précédents travaux sur les machines parallèles, les systèmes multiprocesseurs sur puce (MPSoC) représentent une solution intéressante et prometteuse. Dans de précédents travaux à cette thèse, la cible technologique pour développer de tels systèmes était les FPGA. Or les résultats ont montré les limites de cette cible en terme de ressource matérielles et en terme de performance (vitesse notamment). Ce constat nous amène à changer de cible c’est-à-dire à passer sur cible ASIC nécessitant ainsi de retravailler profondément l’architecture et les IPs qui existaient autour de la méthode existante (appelée HNCP, pour Homogeneous Network of Communicating Processors). Afin de bénéficier de la performance offerte par la cible ASIC, les systèmes multiprocesseurs proposés s’appuient sur la flexibilité de son architecture. Combinés à des squelettes de parallélisation facilitant la programmabilité de l’architecture, les circuits proposés permettent d’offrir des systèmes supportant le portage en temps réels de différentes classes d’algorithme de traitement d’images. Le résultat de ce travail a abouti à la fabrication d’un circuit intégré à base d’un seul processeur et de ses périphériques en technologie ST CMOS 65nm dont la surface est d’environ 1 mm² et à la définition de 2 architectures multiprocesseurs flexibles basées sur le concept des squelettes de parallélisation (une architecture de 16 coeurs de processeur en technologie ST CMOS 65 nm et une deuxième architecture de 64 coeurs de processeur en technologie ST CMOS FD-SOI 28 nm). / Smart sensors today require processing components with sufficient power to run algorithms at the rate of these high-performance image sensors, while maintaining low power consumption. Monoprocessor systems are no longer able to meet the requirements of this field. Thus, thanks to technological advances and based on previous works on parallel computers, multiprocessor systems on chip (MPSoC) represent an interesting and promising solution. Previous works around this thesis have used FPGA as technological target. However, results have shown the limits of this target in terms of hardware resources and in terms of performance (speed in particular). This observation leads us to change the target from FPGA to ASIC. This migration requires deep rework at the architecture level. Particularly, existing IPs around the method (called HNCP for Homogeneous Network of Communicating Processors) have to be revisited. To take advantage of the performance offered by the ASIC target, proposed multiprocessor systems are based on the flexibility of its architecture. Combined with parallel skeletons that ease programmability of the architecture, the proposed circuits allow to offer systems that support various real-time image processing algorithms. This work has led to the fabrication of an integrated circuit based on a single processor and its peripheral using ST CMOS 65nm technology with an area around 1 mm². Moreover, two flexible multiprocessor architectures based on the concept of parallel skeletons have been proposed (a 16 cores 65 nm CMOS multiprocessors and a 64 cores 28 nm FD-SOI CMOS multiprocessors).

The critical role of language pronunciation in communicative competence is significant, especially for second language learners. Despite renewed awareness of the importance of articulation, it remains a challenge for instructors to handle the pronunciation needs of language learners. There are relatively scarce pedagogical tools for pronunciation teaching and learning, such as inefficient, traditional pronunciation instructions like listening and repeating. Recently, electronic visual feedback (EVF) systems (e.g., medical ultrasound imaging) have been exploited in new approaches in such a way that they could be effectively incorporated in a range of teaching and learning contexts. Evaluation of ultrasound-enhanced methods for pronunciation training, such as multimodal methods, has asserted that visualizing articulator’s system as biofeedback to language learners might improve the efficiency of articulation learning. Despite the recent successful usage of multimodal techniques for pronunciation training, manual works and human manipulation are inevitable in many stages of those systems. Furthermore, recognizing tongue shape in noisy and low-contrast ultrasound images is a challenging job, especially for non-expert users in real-time applications. On the other hand, our user study revealed that users could not perceive the placement of their tongue inside the mouth comfortably just by watching pre-recorded videos. Machine learning is a subset of Artificial Intelligence (AI), where machines can learn by experiencing and acquiring skills without human involvement. Inspired by the functionality of the human brain, deep artificial neural networks learn from large amounts of data to perform a task repeatedly. Deep learning-based methods in many computer vision tasks have emerged as the dominant paradigm in recent years. Deep learning methods are powerful in automatic learning of a new job, while unlike traditional image processing methods, they are capable of dealing with many challenges such as object occlusion, transformation variant, and background artifacts. In this dissertation, we implemented a guided language pronunciation training system, benefits from the strengths of deep learning techniques. Our modular system attempts to provide a fully automatic and real-time language pronunciation training tool using ultrasound-enhanced augmented reality. Qualitatively and quantitatively assessments indicate an exceptional performance for our system in terms of flexibility, generalization, robustness, and autonomy outperformed previous techniques. Using our ultrasound-enhanced system, a language learner can observe her/his tongue movements during real-time speech, superimposed on her/his face automatically.

No description available.