ABSTRACT. Human activity recognition has been an important area of computer vision research. In this paper, we present real-time hardware implementation for action recognition with HON4D features, which outperform the methods relying on skeleton detectors. Our proposed circuit adopts sliding histogram, and several approximate techniques to reduce computation and speed up feature extraction. Furthermore, using sliding histogram allows continuous classification without video segmentation in advance.

ABSTRACT. Support Vector Machine (SVM) classifier with Histogram of Oriented Gradients (HOG) features is one of the most popular techniques used in vehicle detection in recent years. The computing time of SVM is a main hurdle to get real time vehicle detection which is important for Advanced Driver Assistance Systems (ADAS) applications. One of the effective ways to reduce the computing complexity of SVM is to reduce the dimension of HOG features. In this paper, we examine the effect of the number of HOG bins on the vehicle detection and the symmetric characteristics of HOG features of vehicles. And we successfully demonstrate the speed-up of SVM classifier for vehicle detection by about three times while maintaining the detection performance.

ABSTRACT. In the future robotic applications, Robot requires the ability not only to interact with people but also to learn incrementally and quickly to adapt to ever-changing environment. Therefore, we proposed an incremental action learning system for this future requirement. The proposed system can continuously learn new actions quickly with robust performance and less effort

ABSTRACT. In this paper, we present an approach for automatically convert images from 2D to 3D. The algorithm uses a color + depth dataset to estimate a depth map of a query color image by searching structurally similar images in the dataset and fusing them. Our experimental results indicate that the inclusion of a retinex based stage for the query image and the dataset images improves the performance of the system on commonly-used databases and for different image descriptors.

ABSTRACT. Wireless sensor networks (WSN) have plenty potential in many monitoring applications like environmental parameters or building and home automation scenarios including, gas, temperature, humidity, presence motion detection or even power grid analyzer. Recent advances in technology, IC manufacturing, and networking utilizing Bluetooth communications have managed to reduce the total power requirements of Bluetooth nodes to the newest Bluetooth Low Energy (BLE) specification. Such nodes can perform measurement and transmission functions for a long time and are supposed to play a big role in the deployment of the IoT (Internet of Things). Our aim is to create a system based on BLE in order to acquire parameters that supply information to a central in-home controller which allows performing a remote device control and monitoring system for home environments.

ABSTRACT. ZigBee Alliance proposed ZigBee IP standard to support wireless sensor mesh networks. However, the MTU of IEEE 802.15.4 MAC frame is only 127-byte. For the architecture of IPv6 over 802.15.4, the ratio of frame overhead is very high and significantly reduces the ratio of carried data. This paper adopts an error detection mechanism with ultra-low overhead to efficiently improve the ratio of carried data. This mechanism utilizes few error checking bits instead of complete CRC-16 checksum and uses chained computing mechanism to compensate and improve the error detection ability. Finally, a mathematical analysis model is constructed to analyze and demonstrate the error detection performance of this mechanism.

ABSTRACT. The cloud is playing a very important role in wireless sensor network, crowd sensing and IoT data collection and processing. However, current cloud solutions lack of some features that hamper the innovation a number of other new services. We propose a cloud solution that provides these missing features as multi-cloud and device multi-tenancy relying in a whole different fully distributed paradigm, the actor model.

ABSTRACT. One of important and difficult issues in power saving is how to determine the most appropriate sleep window timing and size for unpredictable non-real-time traffic. This paper proposes an almost overhead-free dynamic sleep window adjusting policy (D-SWAP) for mobile communication systems. Based on monitoring downlink traffic patterns, it adopts the concept of round-trip-time (RTT) estimation in TCP to enable base station and mobile station to autonomously adjust the sleep window timing and size. The performance evaluation shows that the proposed D-SWAP can concurrently satisfy the requirements of the trade-off problem between power saving and packet delay.

ABSTRACT. Preparation of official examinations and oppositions is a difficult task in which new technologies related to mobile Web 2.0 play an important role, because they can improve the learning methodology. A collaborative platform based on a smartphone application was developed for helping students to prepare an official examination which, in this concrete case, is the preparation of the exam for obtaining the accreditation for touristic guides of the Agència Valenciana del Turisme. The app enables the users to collaborate in the improving the platform sending their own questions and opinions about the development. The gamification of the process helps students increase their motivation to study. The application Turisme GVA had a very good reception with a high amount of downloads, also supported with opinions from users and real collaboration to improve the functioning and interface. All this interaction and exchange of information, complemented with final surveys and statistics, revealed a good level of satisfaction from users in the usage of smartphones to increase students’ motivation and their degree of involvement for reaching a common goal. This project is a pioneer in the introduction and usage of new technologies in e-learning and mobile education.

ABSTRACT. Augmented reality is becoming more and more popular due to the countless number of practical applications. A key element is the understanding of the scene and the involved human activities to be able to offer a rich interaction with the world via virtual actions and elements. For this purpose, a new vision-based human-action recognition module has been developed to be integrated with the new generation of augmented reality devices, which is based on Spatio-Temporal Interest Points and the Citation K-Nearest Neighbor classifier.

ABSTRACT. Syntax analysis in Spanish language is necessary for university entrance examination. Secondary students should control this subject, which is difficult to teach with conventional educational tools. This paper describes a mobile devices application for defining linguistic structure above the word level, i.e. syntax analysis. The application was designed following the official recommendations for this examination. The advantage of this application is that it allows following the process instead of only observing the final result. The application, which is a pioneer in this field, was tested with teachers and students in their last year of secondary education. Conclusions determined that the usage of this tool improve their understanding of the subject and prepare them adequately for their examinations. This kind of self-assessment application could be easily exportable to other subjects, improving students’ results.

ABSTRACT. Multimedia advertising content needs to be assessed before campaign to assure its future success. Content consumption techniques are used to identify user’s interest through a determined content, but it needs effective validation. This paper describes the design of a two-fold analysis of video consumption implemented on EPSIS platform to measure users’ interest. The platform consists in a web interface for video visualization which executes an algorithm for automatic validation of user’s consumption, and a mobile application which subsequently enables users to assess the video content previously visualized. The video portfolio featured in our platform corresponds to advertising content and social media, which has special characteristics of length and complete meaning and is easier to be assessed than longer contents such as documentary or films. After developing subjective tests with a collection of observers, the mobile application allows to demonstrate the validation of the algorithm. The platform succeeds in validation of advertising contents reducing time and cost derived from the evaluation process.

ABSTRACT. In this paper, we design a service platform for dynamic provisioning of SEP compliant energy services. The designed platform is to lead energy market participation of 3rd party energy service providers who have solutions for integrated management of energy generation, consumption, and saving. The concept of open energy service gateway is introduced. The system architecture for dynamic provisioning of energy services is explained. Finally, the operations of each component defined in the designed system are described.

ABSTRACT. In this paper, we design a building energy management system for integrated management of energy generation, consumption, storing, and trading to maintain balance between energy demand and supply in customer buildings. The term “Transactive Energy” defined by the GridWise is introduced to help clear understanding of the designed system. From the main idea of the Transactive Energy, we design an energy information network and system functions based on EI and EMIX. Finally, we explain how the designed system works on logical energy information networks by showing service invocation procedures between its components.

ABSTRACT. In order to develop the next generation of gaze tracking system that can operate under visible lighting conditions, the difficulty of estimating limbus circle in real time needs to be solved. The iris model-based approaches excel the feature-based ones, but high computation complexity remains a problem. This paper presents an advanced iris model-based matching algorithm which adopts particle swarm optimization (PSO) to improve the overall performance. As a result, the developed system achieves high accuracy and the objective of 30 frames.

ABSTRACT. This paper presents design details of a Virtual Platform (VP) which is used for developing real time hardwired face recognition system. VP is promising technology to develop complex HW/SW system because it enables the simultaneous development of hardware and software. We have implemented a Virtual Platform and we used it for developing FPGA based face recognition system. The hardware portion is modeled with Transaction Level Model (TLM) in early phase of development. After hardware have been implemented, TLM model is substituted with real hardware system. The software is simultaneously developed with TLM hardware model and SW/HW integrity verification is seamlessly done by substituting TLM model with FPGA based real hardware. TLM model was efficient because it was much faster than RTL model and it can be seamlessly interfaced with SW development environment of virtual platform. FPGA based face recognition system was fully verified using application software running on virtual platform.

ABSTRACT. Face recognition system plays an important role in many applications such as surveillance, biometrics and security. In this paper, we present a FPGA implementation of real-time face recognition engine consists of face detection, eye detection and feature extraction/verification module. We apply Self-Quotient Image (SQI) to reduce illumination effect. Haar-AdaBoost trained classifiers are applied for face and eye detection. Linear Discriminant Analysis (LDA) and Discrete Cosine Transform (DCT) are used for face recognition. The complete system is implemented in a Xilinx Virtex-7 device. Experimental result shows that our face recognition system operates over 30 fps at 50MHz frequency for real-time applications.

ABSTRACT. This paper presents a wireless power transfer system for mobile devices. To achieve high efficiency in the receiver circuit, an optimal switching frequency is required. Through experimental analysis, we suggest the optimal efficiency of a receiver circuit. The switching frequency of this receiver circuit can vary from 500 kHz to 13.56 MHz. The received power has a range of 0 to 5 W, and it has a maximum efficiency of 81%. This receiver circuit was fabricated with a 0.18 µm BCD process, and it occupies an area of 2.5 mm × 2.5 mm.

ABSTRACT. This paper presents the development of an image data acquisition system that uses a multi-function front camera in automotive applications. The multi-function front camera detects lanes, vehicles, traffic signals, objects, and pedestrians,and the system signals the driver about these obstacles. The image data acquisition system stores and synchronizes both the recognition results and the vehicular information.

ABSTRACT. In this paper, an OS level power estimator based on monitored system events for multimedia mobile devices is presented. The OS level power estimator periodically obtains significant-events count values and calculates power-consumption estimations through mathematical models. We have implemented this power estimator in a Linux kernel and evaluated it while running a video decoder application on an embedded development platform. Experiments results show that the accuracy of power estimations is high with regard to the real measured power consumption, i.e., the error between them is less than 1% in average.

ABSTRACT. This article presents the hardware implementation of the speech recognition for real time performance and high-level accuracy. The stand-alone speech recognizer should simultaneously achieve the requirements, which are the low-latency performance and the low-power dissipation in an environment that cannot connect to the network. So, we made a speech recognizer as the hardware accelerator based on the hidden markov model for reducing the load of the system processor without the cloud computing. Our overall design has the fully hardwired operation flow from the generation of the speech feature to the generation of the recognized words. Our design showed low-latency performance as the real time factor of 0.4 ~ 0.5 on FPGA, which operates at 100MHz operating frequency and uses the resource of 10%.

ABSTRACT. Abstract — In the embedded system, the memory profiling of an application is very critical and important to solve the problem. The memory usage is usually measured on the basis of process-level granularity. However, the process-level memory usage gives a lack of information to analyze an application, which consists of big processes with many threads. If a specific thread in the process has memory problems like memory leak or large memory consumption, then this may lead to a malfunction of an application due to the waste of an unnecessary memory by the problematic thread. This paper gives an efficient framework to get a memory usage of a process with thread level granularity and we called it Memory Usage Per Thread(MUPT). This framework would provide memory allocation information in terms of page_fault for a given multi-threaded process with thread granularity and results are expressed using proc interface. Thus, MUPT provides a mechanism to debug or monitor the thread memory usage and guide an application user to detect memory leak or large amount of incorrect memory allocation.

ABSTRACT. This paper proposes a user immersive augmented reality (AR) system which utilizes three kinds of computer vision technology: projector-camera-based spatial AR, hand-gesture-based user interaction, and mobile AR. The proposed system augments 2D virtual contents on a tabletop using the projector-camera-based AR, and a user can interact with the augmented 2D contents by using natural hand gestures and a physical object by using smart phone. Also, the proposed system further enhances user-immersion by adopting the mobile AR, which augments 3D virtual contents on the physical object when user sees the object through smart phone’s camera. This paper verifies the effectiveness of the proposed AR system through implementing a car driving application.

ABSTRACT. Through various cases of inverter layout, the change in the propagation delay time (tPD) in the ring oscillator that consists of inverters can be analyzed. In this paper, an inverter layout technique for tPD minimization is presented. Through the case-by-case layout, to reduce the tPD, we propose that layout engineers should reduce the input and output node length. The proposed technique post-simulated in a 0.18um CMOS process achieves maximum 7.318% reduced tPD compared to the basic inverter layout.

ABSTRACT. In various situations surrounding noise can be disturbing or stressful. Active Noise Control (ANC) is an effective technology for noise reduction that works as a good complement of the passive isolation offered by traditional headphones. Digital ANC systems are progressively finding their way into consumer electronic applications, in replacement of conventional ANC devices based on analog circuits. This paper introduces a novel digital ANC architecture based a new class of audio sensors: Pulse-Density Modulation (PDM) MEMS digital microphones. The benefits of the proposed ANC system are highlighted, and an experimental validation is described.

ABSTRACT. This paper proposes a noble around view system based on moving object detection algorithm using background subtraction for safe bus getting on-and-off. Recently, vehicles tend to equip with several types of cameras on them as standard option for safety. Since full sized vehicles like buses confront with more moving persons around them comparing with small sized cars, they may need around view monitoring system instead of single rear camera to enlarge driver’s view angle. The proposed system provides drivers with environment information around the vehicle for safe getting on-and-off. This algorithm requires high processing speed rather than high detection rate because on-line detection and action are important to avoid possible collision with them around vehicle. The proposed algorithm detects the locations of objects using information of object’s motion obtained by applying a simple background subtraction method which compares the consecutive frames. The experiment which uses real movies in embedded system and PC confirms the effectiveness of the object detection algorithm.

ABSTRACT. This paper presents an overview of the European project HBB4ALL and tries to answer three questions: how access services can be provided by means of HbbTV applications, why connected TV features become useful for this purpose and how to get fully satisfying access services for the citizens.

ABSTRACT. We consider the introduction of three different methods to signalize which data packets are used to generate each repair packet in systematic coding schemes employing flexible connection setups: a first one based on a mask, a second one on explicitly pointing out the links between the two packet streams, and a third one on sharing the signaling load between both flows. These methods have been evaluated in terms of the bandwidth required to put them into practice. The evaluation results show that the explicit method introduces less bandwidth, whereas the one sharing the load handles it more smartly.

ABSTRACT. In this paper, we consider three-level hierarchical 64-QAM with scalable multi-view video coding (SMVC) for seamless free-viewpoint TV (FTV) service. Performance evaluation results show that our proposed scheme achieves improved average peak signal-to-noise ratio (PSNR), compared to SMVC with conventional 64-QAM.

ABSTRACT. In free-viewpoint TV system considering hybrid transmission through heterogeneous networks, our proposed scheme using Zadoff-Chu sequence can improve synchronization performance of this system.

ABSTRACT. Fog Computing is a recent computing paradigm that is extending cloud computing towards the edge of network. Due to its proximity to end-users, dense geographical distribution, open platform and support for high mobility, Fog Computing platforms can provide services with reduced latency and improved QoS. Thus it is becoming an important enabler for consumer centric Internet of Things based applications and services that require real time operations e.g. connected vehicles, smart road intersection management and smart grid. The paper discusses one such architecture for connected vehicles with Road Side Units (RSUs) and M2M gateways including the Fog Computing Platform. M2M data processing with semantics, discovery and management of connected vehicles are briefly discussed as consumer centric IoT services enabled by the distinct characteristics of Fog Computing.

ABSTRACT. This article describes our approach for facilitating the interaction among devices in IoT environments. Our solution provides mechanisms to complement current IoT ontologies with device language annotations to facilitate device communication. This is our first step towards comprehensive user to environment communication that would bring the Internet of Me concept.

ABSTRACT. This paper presents an automatic liquid level detection algorithm for a beverage refilling service in a restaurant. The automatic detection of liquid level provides the amount of residual of customer’s beverage using a general-purpose surveillance camera deployed in a restaurant. The proposed system consists of three modules i) extraction of a cup and beverage region using color-based region segmentation by K-means clustering, ii) edge detection of the cup and beverage region, and iii) liquid level estimation based on edge detection. Using a general-purpose CCTV camera and a set of relevant image processing algorithms, the proposed system can be one of important applications of internet of things (IoT) service.

ABSTRACT. This paper proposes a software framework and hardware architecture to exchange the binaries and profiles from smartphones to IoT devices and reprogram devices with smartphones. This scheme, when included in a system of smartphones interconnected with locally deployed IoT devices in wired or wireless manners, can be used to implement a practical updater.