Traffic sign recognition is an important problem for autonomous cars and driver assistance systems. With recent developments in the field of machine learning, high performance can be achieved, but typically at a large computational cost. This thesis aims to investigate the relation between classification accuracy and computational complexity for the visual recognition problem of classifying traffic signs. In particular, the benefits of partitioning the classification problem into smaller sub-problems using prior knowledge in the form of shape or current region are investigated. In the experiments, the convolutional neural network (CNN) architecture MobileNetV2 is used, as it is specifically designed to be computationally efficient. To incorporate prior knowledge, separate CNNs are used for the different subsets generated when partitioning the dataset based on region or shape. The separate CNNs are trained from scratch or initialized by pre-training on the full dataset. The results support the intuitive idea that performance initially increases with network size and indicate a network size where the improvement stops. Including shape information using the two investigated methods does not result in a significant improvement. Including region information using pretrained separate classifiers results in a small improvement for small complexities, for one of the regions in the experiments. In the end, none of the investigated methods of including prior knowledge are considered to yield an improvement large enough to justify the added implementational complexity. However, some other methods are suggested, which would be interesting to study in future work.

Deep learning has in recent years come to dominate the previously separate fields of research in machine learning, computer vision, natural language understanding and speech recognition. Despite breakthroughs in training deep networks, there remains a lack of understanding of both the optimization and structure of deep networks. The approach advocated by many researchers in the field has been to train monolithic networks with excess complexity, and strong regularization --- an approach that leaves much to desire in efficiency. Instead we propose that carefully designing networks in consideration of our prior knowledge of the task and learned representation can improve the memory and compute efficiency of state-of-the art networks, and even improve generalization --- what we propose to denote as structural priors. We present two such novel structural priors for convolutional neural networks, and evaluate them in state-of-the-art image classification CNN architectures. The first of these methods proposes to exploit our knowledge of the low-rank nature of most filters learned for natural images by structuring a deep network to learn a collection of mostly small, low-rank, filters. The second addresses the filter/channel extents of convolutional filters, by learning filters with limited channel extents. The size of these channel-wise basis filters increases with the depth of the model, giving a novel sparse connection structure that resembles a tree root. Both methods are found to improve the generalization of these architectures while also decreasing the size and increasing the efficiency of their training and test-time computation. Finally, we present work towards conditional computation in deep neural networks, moving towards a method of automatically learning structural priors in deep networks. We propose a new discriminative learning model, conditional networks, that jointly exploit the accurate representation learning capabilities of deep neural networks with the efficient conditional computation of decision trees. Conditional networks yield smaller models, and offer test-time flexibility in the trade-off of computation vs. accuracy.

Les Réseaux de Neurones Convolutifs profonds (CNNs) ont connu un large succès au cours de la dernière décennie, devenant un standard de la vision par ordinateur. Ce succès s’est fait au détriment d’un large coût de calcul, où le déploiement des CNNs reste une tâche ardue surtout sous des contraintes de temps réel.Afin de rendre ce déploiement possible, la littérature exploite le parallélisme important de ces algorithmes, ce qui nécessite l’utilisation de plate-formes matérielles dédiées. Dans les environnements soumis à des contraintes de consommations énergétiques, tels que les nœuds des caméras intelligentes, les cœurs de traitement à base de FPGAs sont reconnus comme des solutions de choix pour accélérer les applications de vision par ordinateur. Ceci est d’autant plus vrai pour les CNNs, où les traitements se font naturellement sur un flot de données, rendant les architectures matérielles à base de FPGA d’autant plus pertinentes. Dans ce contexte, cette thèse aborde les problématiques liées à l’implémentation des CNNs sur FPGAs. En particulier, ces travaux visent à améliorer l’efficacité des implantations grâce à deux principales stratégies d’optimisation; la première explore le modèle et les paramètres des CNNs, tandis que la seconde se concentre sur les architectures matérielles adaptées au FPGA. / Deep Convolutional Neural Networks (CNNs) have become a de-facto standard in computer vision. This success came at the price of a high computational cost, making the implementation of CNNs, under real-time constraints, a challenging task.To address this challenge, the literature exploits the large amount of parallelism exhibited by these algorithms, motivating the use of dedicated hardware platforms. In power-constrained environments, such as smart camera nodes, FPGA-based processing cores are known to be adequate solutions in accelerating computer vision applications. This is especially true for CNN workloads, which have a streaming nature that suits well to reconfigurable hardware architectures.In this context, the following thesis addresses the problems of CNN mapping on FPGAs. In Particular, it aims at improving the efficiency of CNN implementations through two main optimization strategies; The first one focuses on the CNN model and parameters while the second one considers the hardware architecture and the fine-grain building blocks.

<p>In recent years, deep learning models have become popular in the real-time embedded application, but there are many complexities for hardware deployment because of limited resources such as memory, computational power, and energy. Recent research in the field of deep learning focuses on reducing the model size of the Convolution Neural Network (CNN) by various compression techniques like Architectural compression, Pruning, Quantization, and Encoding (e.g., Huffman encoding). Network pruning is one of the promising technique to solve these problems.</p> <p>This thesis proposes methods to prune the convolution neural network (SqueezeNet) without introducing network sparsity in the pruned model. </p> <p>This thesis proposes three methods to prune the CNN to decrease the model size of CNN without a significant drop in the accuracy of the model.</p> <p>1: Pruning based on Taylor expansion of change in cost function Delta C.</p> <p>2: Pruning based on L₂ normalization of activation maps.</p> <p>3: Pruning based on a combination of method 1 and method 2.</p><p>The proposed methods use various ranking methods to rank the convolution kernels and prune the lower ranked filters afterwards SqueezeNet model is fine-tuned by backpropagation. Transfer learning technique is used to train the SqueezeNet on the CIFAR-10 dataset. Results show that the proposed approach reduces the SqueezeNet model by 72% without a significant drop in the accuracy of the model (optimal pruning efficiency result). Results also show that Pruning based on a combination of Taylor expansion of the cost function and L₂ normalization of activation maps achieves better pruning efficiency compared to other individual pruning criteria and most of the pruned kernels are from mid and high-level layers. The Pruned model is deployed on BlueBox 2.0 using RTMaps software and model performance was evaluated.</p><p></p>

In the development of advanced driver assistance systems, computer vision problemsneed to be optimized to run efficiently on embedded platforms. Convolutional neural network(CNN) accelerators have proven to be very efficient for embedded camera platforms,such as the ones used for automotive vision systems. Therefore, the focus of this thesisis to evaluate the efficiency of a CNN on a future embedded heterogeneous processingdevice. The memory size in an embedded system is often very limited, and it is necessary todivide the input into multiple tiles. In addition, there are power and speed constraintsthat needs to be met to be able to use a computer vision system in a car. To increaseefficiency and optimize the memory usage, different methods for CNN layer fusion areproposed and evaluated for a variety of tile sizes. Several different layer fusion methods and input tile sizes are chosen as optimal solutions,depending on the depth of the layers in the CNN. The solutions investigated inthe thesis are most efficient for deep CNN layers, where the number of channels is high.

<p>Abstract</p><p>Title: Foreign news coverage. Does the media influence the action in international crises? (Utrikesbevakning. Påverkar media agerandet i internationella kriser?)</p><p>Number of pages: 39</p><p>Author: Charlotta Andrésson</p><p>Tutor: Professor Lowe Hedman</p><p>Course: Media and Communication Studies C</p><p>University: Division of Media and Communication, Department of Information Science, Uppsala University.</p><p>Date of submission: 2007-01-03, autumn term of 2006</p><p>Purpose/Aim</p><p>The purpose of the essay is partly to examine if foreign news coverage influence the political agenda setting and the incentives of the public’s willingness to give charity for humanitarian crises. It is also to answer if the media are responsible for the possible consequences of the news coverage. My main questions at issue are:</p><p>1. Does the foreign news coverage influence the political agenda setting and the incentives of the public’s willingness to give charity for humanitarian crises?</p><p>2. Is media responsible for the possible consequences of their foreign news coverage?</p><p>I also ask a question at issue in a research of Swedish foreign news coverage in my essay to get a clearer picture of the foreign news coverage:</p><p>3. How does Swedish foreign news coverage relate to prior research?</p><p>Method:</p><p>The second chapter of the essay is a literature research of news selection and news values. The third chapter of the essay is a research of media’s influence on the world politics and humanitarian aid. These two chapters are literature studies based on prior research, theories and debates. The fourth chapter is an empirical study of a news programme on a Swedish TV-channel during a five months period between 2004 and 2005. My interest in the empirical study was to examine how the material was divided geographically and as regards contents. The fifth and sixth chapter of the essay consists of an analysis and a discussion.</p><p>Main results:</p><p>As my main result I concluded that the media influence the political agenda setting and the the incentives of the public’s willingness to give charity for humanitarian crises. The media throw light upon which crises that should be given priority to. The theories for news selection and news value agrees with the result of my research of Swedish foreign news coverage. I also concluded that the media alone was not responsible for the possible consequences for their foreign news coverage but that they are the premier channel of information about the world for most people.</p><p>Keywords: Foreign news, news selection and news value, CNN-effect, Media and political agenda setting, Media influence of humanitarian aid.</p>

Abstract Title: Foreign news coverage. Does the media influence the action in international crises? (Utrikesbevakning. Påverkar media agerandet i internationella kriser?) Number of pages: 39 Author: Charlotta Andrésson Tutor: Professor Lowe Hedman Course: Media and Communication Studies C University: Division of Media and Communication, Department of Information Science, Uppsala University. Date of submission: 2007-01-03, autumn term of 2006 Purpose/Aim The purpose of the essay is partly to examine if foreign news coverage influence the political agenda setting and the incentives of the public’s willingness to give charity for humanitarian crises. It is also to answer if the media are responsible for the possible consequences of the news coverage. My main questions at issue are: 1. Does the foreign news coverage influence the political agenda setting and the incentives of the public’s willingness to give charity for humanitarian crises? 2. Is media responsible for the possible consequences of their foreign news coverage? I also ask a question at issue in a research of Swedish foreign news coverage in my essay to get a clearer picture of the foreign news coverage: 3. How does Swedish foreign news coverage relate to prior research? Method: The second chapter of the essay is a literature research of news selection and news values. The third chapter of the essay is a research of media’s influence on the world politics and humanitarian aid. These two chapters are literature studies based on prior research, theories and debates. The fourth chapter is an empirical study of a news programme on a Swedish TV-channel during a five months period between 2004 and 2005. My interest in the empirical study was to examine how the material was divided geographically and as regards contents. The fifth and sixth chapter of the essay consists of an analysis and a discussion. Main results: As my main result I concluded that the media influence the political agenda setting and the the incentives of the public’s willingness to give charity for humanitarian crises. The media throw light upon which crises that should be given priority to. The theories for news selection and news value agrees with the result of my research of Swedish foreign news coverage. I also concluded that the media alone was not responsible for the possible consequences for their foreign news coverage but that they are the premier channel of information about the world for most people. Keywords: Foreign news, news selection and news value, CNN-effect, Media and political agenda setting, Media influence of humanitarian aid.

This thesis analyzes the role of the global media in world politics. The global media as a major source of information performs many functions in world politics. Due to the technological innovations both the global media and world politics has extended their scope and content. It is the aim of this thesis to explore to what extent the global media and world politics changes and to what extent the global media affect world politics. Moreover, this thesis aims to analyze the actors that play a significant role in the relation of global media and world politics. This thesis will also question the importance and effects of global media in world politics through the examples of Iraq War of 2003. This thesis argues that global media are the important actor of world politics by their agenda-setting, impediment and accelerant effects which influences public opinion. The aim of this thesis is to question the power of the global media on public opinion through the existential media structures, while showing the effects of global media by the examples of Iraq War of 2003.

With self-driving cars on the horizon, vehicle autonomy and its problems is a hot topic. In this study we are using convolutional neural networks to make a robot car avoid obstacles. The robot car has a monocular camera, and our approach is to use the images taken by the camera as input, and then output a steering command. Using this method the car is to avoid any object in front of it. In order to lower the amount of training data we use models that are pretrained on ImageNet, a large image database containing millions of images. The model are then trained on our own dataset, which contains of images taken directly by the robot car while driving around. The images are then labeled with the steering command used while taking the image. While training we experiment with using different amounts of frozen layers. A frozen layer is a layer that has been pretrained on ImageNet, but are not trained on our dataset. The Xception, MobileNet and VGG16 architectures are tested and compared to each other. We find that a lower amount of frozen layer produces better results, and our best model, which used the Xception architecture, achieved 81.19% accuracy on our test set. During a qualitative test the car avoid collisions 78.57% of the time.

In recent years, development of Convolutional Neural Networks has enabled high performing semantic segmentation models. Generally, these deep learning based segmentation methods require a large amount of annotated data. Acquiring such annotated data for semantic segmentation is a tedious and expensive task. Within machine learning, active learning involves in the selection of new data in order to limit the usage of annotated data. In active learning, the model is trained for several iterations and additional samples are selected that the model is uncertain of. The model is then retrained on additional samples and the process is repeated again. In this thesis, an active learning framework has been applied to road segmentation which is semantic segmentation of objects related to road scenes. The uncertainty in the samples is estimated with Monte Carlo dropout. In Monte Carlo dropout, several dropout masks are applied to the model and the variance is captured, working as an estimate of the model’s uncertainty. Other metrics to rank the uncertainty evaluated in this work are: a baseline method that selects samples randomly, the entropy in the default predictions and three additional variations/extensions of Monte Carlo dropout. Both the active learning framework and uncertainty estimation are implemented in the thesis. Monte Carlo dropout performs slightly better than the baseline in 3 out of 4 metrics. Entropy outperforms all other implemented methods in all metrics. The three additional methods do not perform better than Monte Carlo dropout. An analysis of what kind of uncertainty Monte Carlo dropout capture is performed together with a comparison of the samples selected by baseline and Monte Carlo dropout. Future development and possible improvements are also discussed.