Multi-Sensor Motion Fusion Using Deep Neural Network Learning

\n Hand pose estimation is a challenging problem in computer vision with a wide range of applications, especially in human-computer interface. With the development of inexpensive consumer-level depth cameras and the evolution on deep learning techniques, the current state-of-art in the problem is continuously developing and several new methods have been proposed in recent years. Those methods are mostly data-driven and reach good results in standard datasets such as NYU, ICVL and HANDS17. An application that would benefit from the use of computer vision techniques is hand occupational therapy. In chronic diseases like rheumatoid arthritis (RA), the evaluation of the hand functional state is fundamental for the treatment and prevention of finger deformities. One of the procedures for deformity diagnosis is the measurement of movement angles i.e. flexion/extension and abduction/addution, made using goniometers in a process that can be time-consuming and invasive for the patient. The main proposal of this PhD is to fill a gap in the literature by proposing and evaluating the viability of using a framework composed of a 3D low-cost sensor and a 3D hand pose estimation state-of-art method for automatic assessment of rheumatoid arthritis patients. Given depth maps as input, our framework estimates 3D hand joint positions using a deep convolutional neural network. The proposed pose estimation algorithm can be executed in real-time, allowing users to visualise 3D skeleton tracking results at the same time as the depth images are acquired. Once 3D joint poses are obtained, our framework estimates flexion/extension and abduction/adduction angles by applying computational geometry oper- ations. The absence of public datasets with RA patients in the literature makes the estimation of hand poses of patients a challenge for computer vision data-driven methods. We therefore proposed a protocol to acquire new data from groups of patients and control. We performed experiments of identification of RA patients and control sets and also performed comparison with goniometer data. Results show that a method based on Fourier descriptors is able to perform automatic discrimination of hands with Rheumatoid Arthritis (RA) and healthy pa- tients. The angle between joints can be used as an indicative of current movement capabilities and function. The acquisition is much easier, non-invasive and patient-friendly, significantly reducing the evaluation time and offering real-time data for the dynamic movement.\n

Hand pose estimation from 3D data is a key challenge in computer vision as well as an essential step for human–computer interaction. A lot of deep learning-based hand pose estimation methods have made significant progress but give less consideration to the inner interactions of input data, especially when consuming hand point clouds. Therefore, this paper proposes an end-to-end capsule-based hand pose estimation network (Capsule-HandNet), which processes hand point clouds directly with the consideration of structural relationships among local parts, including symmetry, junction, relative location, etc. Firstly, an encoder is adopted in Capsule-HandNet to extract multi-level features into the latent capsule by dynamic routing. The latent capsule represents the structural relationship information of the hand point cloud explicitly. Then, a decoder recovers a point cloud to fit the input hand point cloud via a latent capsule. This auto-encoder procedure is designed to ensure the effectiveness of the latent capsule. Finally, the hand pose is regressed from the combined feature, which consists of the global feature and the latent capsule. The Capsule-HandNet is evaluated on public hand pose datasets under the metrics of the mean error and the fraction of frames. The mean joint errors of Capsule-HandNet on MSRA and ICVL datasets reach 8.85 mm and 7.49 mm, respectively, and Capsule-HandNet outperforms the state-of-the-art methods on most thresholds under the fraction of frames metric. The experimental results demonstrate the effectiveness of Capsule-HandNet for 3D hand pose estimation.

We introduce in this work an efficient approach for audio scene classification using deep recurrent neural networks. An audio scene is firstly transformed into a sequence of high-level label tree embedding feature vectors. The vector sequence is then divided into multiple subsequences on which a deep GRU-based recurrent neural network is trained for sequence-to-label classification. The global predicted label for the entire sequence is finally obtained via aggregation of subsequence classification outputs. We will show that our approach obtains an F1-score of 97.7% on the LITIS Rouen dataset, which is the largest dataset publicly available for the task. Compared to the best previously reported result on the dataset, our approach is able to reduce the relative classification error by 35.3%.

Taiwan, being located on a path in the west Pacific Ocean where typhoons often strike, is often affected by typhoons. The accompanying strong winds and torrential rains make typhoons particularly damaging in Taiwan. Therefore, we aimed to establish an accurate wind speed prediction model for future typhoons, allowing for better preparation to mitigate a typhoon’s toll on life and property. For more accurate wind speed predictions during a typhoon episode, we used cutting-edge machine learning techniques to construct a wind speed prediction model. To ensure model accuracy, we used, as variable input, simulated values from the Weather Research and Forecasting model of the numerical weather prediction system in addition to adopting deeper neural networks that can deepen neural network structures in the construction of estimation models. Our deeper neural networks comprise multilayer perceptron (MLP), deep recurrent neural networks (DRNNs), and stacked long short-term memory (LSTM). These three model-structure types differ by their memory capacity: MLPs are model networks with no memory capacity, whereas DRNNs and stacked LSTM are model networks with memory capacity. A model structure with memory capacity can analyze time-series data and continue memorizing and learning along the time axis. The study area is northeastern Taiwan. Results showed that MLP, DRNN, and stacked LSTM prediction error rates increased with prediction time (1–6 hours). Comparing the three models revealed that model networks with memory capacity (DRNN and stacked LSTM) were more accurate than those without memory capacity. A further comparison of model networks with memory capacity revealed that stacked LSTM yielded slightly more accurate results than did DRNN. Additionally, we determined that in the construction of the wind speed prediction model, the use of numerically simulated values reduced the error rate approximately by 30%. These results indicate that the inclusion of numerically simulated values in wind speed prediction models enhanced their prediction accuracy.

Survey on Deep Neural Networks in Speech and Vision Systems

We propose the detection of humans using millimeter-wave FMCW radar based on time-varying signatures of range-Doppler diagrams using deep recurrent neural networks (DRNNs). Demand for human detection is increasing for security, surveillance, and search and rescue purposes, recently, with a particular focus on urban areas filled with clutter and moving targets. We suggest the classification of targets based on their signatures in range-Doppler plots with time because the signatures can be consecutively observed. We measure five target types: humans, cars, cyclists, dogs, and road clutter using millimeter-wave FMCW radar that transmits fast chirps at 77 GHz. To maximize the classification accuracy using the time-varying range-Doppler signatures of the targets, we investigate and compare the performance of 2-D-deep convolutional neural networks (DCNN), 3-D-DCNN, and DRNN along with 2-D-DCNN. The DRNN plus 2-D-DCNN showed the best performance, and the classification accuracy yields 99%, with the human classification rate of 100%.

Deep unfolding methods design deep neural networks as learned variations of optimization algorithms through the unrolling of their iterations. These networks have been shown to achieve faster convergence and higher accuracy than the original optimization methods. In this line of research, this paper presents novel interpretable deep recurrent neural networks (RNNs), designed by the unfolding of iterative algorithms that solve the task of sequential signal reconstruction (in particular, video reconstruction). The proposed networks are designed by accounting that video frames' patches have a sparse representation and the temporal difference between consecutive representations is also sparse. Specifically, we design an interpretable deep RNN (coined reweighted-RNN) by unrolling the iterations of a proximal method that solves a reweighted version of the l1 - l1 minimization problem. Due to the underlying minimization model, our reweighted-RNN has a different thresholding function (alias, different activation function) for each hidden unit in each layer. In this way, it has higher network expressivity than existing deep unfolding RNN models. We also present the derivative l1 - l1 -RNN model, which is obtained by unfolding a proximal method for the l1 - l1 minimization problem. We apply the proposed interpretable RNNs to the task of video frame reconstruction from low-dimensional measurements, that is, sequential video frame reconstruction. The experimental results on various datasets demonstrate that the proposed deep RNNs outperform various RNN models.

Deep learning hyperparameter optimization: Application to electricity and heat demand prediction for buildings

<p dir="ltr">Recurrent Neural Networks (RNNs) are a major class of Artificial Neural Networks (ANNs). Suitable network architecture is vital for RNNs to achieve high performance. However, designing the architecture of deep RNNs (DRNNs) is a complicated and time-consuming process. Each layer of a DRNN contains numerous hyper-parameters. It is challenging to optimize the hyper-parameters of all layers jointly to achieve the best possible learning performance. Trial-and-error methods for designing DRNNs have been proven to be laborious and highly costly in practice. Therefore, an efficient automatic architecture search technique is needed to design DRNNs.</p><p dir="ltr">Genetic Algorithm (GA) is a popular evolutionary computation based approach. Existing research works already explored the use of GAs in designing ANNs. But, most of the existing approaches focus on the design of fixed-depth ANNs. However, it is not easy to fix the depth of DRNNs in advance, since different problems need DRNNs of different depths. Hence, the potential of GA has to be further explored to evolve the architecture of DRNNs of varying depths.</p><p dir="ltr">The primary goal of this thesis is to develop advanced GA approaches for designing Long Short Term Memory (LSTM) based DRNN architectures of varying depths effectively and efficiently.</p><p dir="ltr">Firstly, this thesis proposes a GA-based algorithm called Two-Stage Surrogate-Assisted GA (TS-SA-GA) with a progressive incremental strategy to design the architecture of LSTM networks of varying depths. The progressive approach can effectively extend well-designed shallow networks to high performing deep networks. Moreover, the new GA-based algorithm adopts newly designed knowledge-driven crossover and mutation operators to identify and repair LSTM network designs affected by inappropriate use of activation functions, thereby significantly reducing the chances for GA to evolve hard-to-train LSTM network architectures. Furthermore, this thesis proposes a two-stage surrogate method to predict the trainability and fitness of LSTM networks. This improves the efficiency and effectiveness of the GA-based algorithm to evolve LSTM networks.</p><p dir="ltr">Secondly, this thesis proposes a new algorithm called Evolutionary Design of LSTM Ensembles (ED-LSTM-Ensemble) based on multi-objective optimization techniques to evolve LSTM networks and ensembles simultaneously to directly utilize the ensemble performance to drive the evolution of base LSTM networks. Additionally, this thesis proposes a connection weight inheritance strategy to evolve LSTM networks and ensembles efficiently and effectively.</p><p dir="ltr">Thirdly, this thesis proposes a new algorithm called Skip Connections through Evolutionary and Differential Architecture search (SCEDA) to design LSTM networks together with appropriate skip connections automatically. Designing LSTM networks together with skip connections is highly challenging. This thesis hybridize gradient-based DAS and evolutionary architecture search to optimize the architecture of LSTM networks with suitable skip connections in a single evolutionary process to improve the efficiency and effectiveness of the search process.</p><p dir="ltr">Finally, this thesis conducts an in-depth empirical analysis of the impact of activation functions on the trainability of LSTM networks based on the concept of Edge Of Chaos. The analysis reveals the strong interrelation between the activation functions across multiple layers of an LSTM network and the trainability of the network. The analysis also highlights the importance of controlling activation functions to avoid generating untrainable LSTM networks. On the basis of the analysis, this thesis proposes a machine learning model to guide the use of activation functions in LSTM networks.</p>

This paper presents an enhanced machine learning based state of charge (SOC) estimation method for a lithium-ion battery using a deep recurrent neural network (DRNN) algorithm. DRNN is suitable for SOC evaluation due to strong computation intelligence and self-learning capabilities. Nevertheless, the performance of DRNN is constrained due to the training accuracy and duration which entirely depends on the appropriate selection of hyper-parameters including hidden layer and hidden neurons. Therefore, firefly algorithm (FA) is employed to find the optimal number for hyper-parameters of DRNN networks. The optimized DRNN based FA algorithm for SOC estimation does not require extensive knowledge about battery chemistry, electrochemical battery model and added filter, rather only needs battery test bench to measure current and voltage. The developed model is tested using two different types of lithium-ion batteries namely lithium nickel manganese cobalt oxide (LiNiMnCoO 2 ) and lithium nickel cobalt aluminum oxide (LiNiCoAlO 2 ). The proposed model is validated by two experimental tests; one with static discharge test and other with pulse discharge test at room temperature. The experimental results indicate the superiority of the DRNN based FA method in comparison with the back-propagation neural network (BPNN) and radial basis function neural network (RBFNN).

Recently the development of deep learning has been propelling the sheer growth of vision and speech applications on lightweight embedded and mobile systems. However, the limitation of computation resource and power delivery capability in embedded platforms is recognized as a significant bottleneck that prevents the systems from providing real-time deep learning ability, since the inference of deep convolutional neural networks (CNNs) and recurrent neural networks (RNNs) involves large quantities of weights and operations. Particularly, how to provide quality-of-services (QoS)-guaranteed neural network inference ability in the multitask execution environment of multicore SoCs is even more complicated due to the existence of resource contention. In this article, we present a novel deep neural network architecture, MV-Net, which provides performance elasticity and contention-aware self-scheduling ability for QoS enhancement in mobile computing systems. When the constraints of QoS, output accuracy, and resource contention status of the system change, MV-Net can dynamically reconfigure the corresponding neural network propagation paths and thus achieves an effective tradeoff between neural network computational complexity and prediction accuracy via approximate computing. The experimental results show that (1) MV-Net significantly improves the performance flexibility of current CNN models and makes it possible to provide always-guaranteed QoS in a multitask environment, and (2) it satisfies the quality-of-results (QoR) requirement, outperforming the baseline implementation significantly, and improves the system energy efficiency at the same time.

Automated image classification is an essential task of the computer vision field. The tagging of images into a set of predefined groups is referred to as image classification. The implementation of computer vision to automate image classification would be beneficial because manual image evaluation and identification can be time-consuming, particularly when there are many images of different classes. Deep learning approaches are proven to overperform existing machine learning techniques in many fields in recent years, and computer vision is one of the most notable examples. The very deep neural network (VDCNN) is a powerful deep learning model for image classification, and this paper examines it briefly using MNIST handwritten digit dataset. This dataset is used to prove the efficacy of very deep neural networks over other deep learning models. The proposed study aims to comprehend the very deep neural network architecture used to accomplish a handwritten digit recognition task. The feasibility of the proposed model is evaluated using mean accuracy, validation accuracy, and standard deviation. The study results of the very deep neural network model are compared to a convolutional neural network and convolutional neural network with batch normalization. According to the results of the comparison study, very deep neural networks achieve high accuracy of 99.1% for a handwritten dataset. The outcome of the proposed work is used to interpret how well a very deep neural network performs when compared to the other two models of deep neural networks. This proposed architecture may be used to automate the classification of handwritten digits dataset.

The recognition of image text sequences has been the subject of long-term research in computer vision. The recognition problem of the shell number image sequence studied in this paper is based on the deep convolutional neural network architecture, and an image recognition algorithm that integrates feature extraction, sequence modeling, and transcription into a unified framework is adopted. The algorithm used has four characteristics: (1) Compared with most algorithms that require separate training and coordination, the method adopted in this paper has end-to-end characteristics. (2) It can handle indefinite length image sequences without involving character segmentation or horizontal-scale normalization. (3) The algorithm is not limited to any predefined vocabulary, and has achieved remarkable performance in both no-lexicon and dictionary-based scene text recognition tasks. (4) The algorithm produces an efficient and much smaller model, which is more practical for real-world scenarios. Using this algorithm for bay number recognition can assist in scene location and provide smart services and application through AI technology.

Text recognition in natural scene images is a challenging problem in computer vision. Different than the optical character recognition (OCR), text recognition in natural scene images is more complex due to variations in text size, colors, fonts, orientations, complex backgrounds, occlusion, illuminations and uneven lighting conditions. In this paper, we propose a segmentation-free method based on a deep convolutional recurrent neural network to solve the problem of cursive text recognition, particularly focusing on Urdu text in natural scenes. Compared to the non-cursive scripts, Urdu text recognition is more complex due to variations in the writing styles, several shapes of the same character, connected text, ligature overlapping, stretched, diagonal and condensed text. The proposed model gets a whole word image as an input without pre-segmenting into individual characters, and then transforms into the sequence of the relevant features. Our model is based on three components: a deep convolutional neural network (CNN) with shortcut connections to extract and encode the features, a recurrent neural network (RNN) to decode the convolutional features, and a connectionist temporal classification (CTC) to map the predicted sequences into the target labels. To increase the text recognition accuracy further, we explore deeper CNN architectures like VGG-16, VGG-19, ResNet-18 and ResNet-34 to extract more appropriate Urdu text features, and compare the recognition results. To conduct the experiments, a new large-scale benchmark dataset of cropped Urdu word images in natural scenes is developed. The experimental results show that the proposed deep CRNN network with shortcut connections outperform than other network architectures. The dataset is publicly available and can be downloaded from <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://data.mendeley.com/datasets/k5fz57zd9z/1</uri> .

We propose a new deep recurrent neural network (RNN) architecture for sequential signal reconstruction. Our network is designed by unfolding the iterations of the proximal gradient method that solves the l 1 -l 1 minimization problem. As such, our network leverages by design that signals have a sparse representation and that the difference between consecutive signal representations is also sparse. We evaluate the proposed model in the task of reconstructing video frames from compressive measurements and show that it outperforms several state-of-the-art RNN models.