2-D complex Gabor filtering has found numerous applications in the fields of computer vision and image processing. Especially, in some applications, it is often needed to compute 2-D complex Gabor filter bank consisting of the 2-D complex Gabor filtering outputs at multiple orientations and frequencies. Although several approaches for fast 2-D complex Gabor filtering have been proposed, they primarily focus on reducing the runtime of performing the 2-D complex Gabor filtering once at specific orientation and frequency. To obtain the 2-D complex Gabor filter bank output, existing methods are repeatedly applied with respect to multiple orientations and frequencies. In this paper, we propose a novel approach that efficiently computes the 2-D complex Gabor filter bank by reducing the computational redundancy that arises when performing the Gabor filtering at multiple orientations and frequencies. The proposed method first decomposes the Gabor basis kernels to allow a fast convolution with the Gaussian kernel in a separable manner. This enables reducing the runtime of the 2-D complex Gabor filter bank by reusing intermediate results of the 2-D complex Gabor filtering computed at a specific orientation. Experimental results demonstrate that our method runs faster than state-of-the-arts methods for fast 2-D complex Gabor filtering, while maintaining similar filtering quality.

In general, the detection of the vehicle license plate is a previous step of license plate recognition and has been actively studied for several decades. In this paper, we propose an algorithm to detect a license plate area of a moving vehicle from a video captured by a fixed camera installed on the road using the Convolution Neural Network (CNN) technology. First, license plate images and non-license plate images are applied to a previously learned CNN model (AlexNet) to extract and classify features. Then, after detecting the moving vehicle in the video, CNN detects the license plate area by comparing the features of the license plate region with the features of the license plate area. Experimental result shows relatively good performance in various environments such as incomplete lighting, noise due to rain, and low resolution. In addition, to protect personal information this proposed system can also be used independently to detect the license plate area and hide that area to secure the public's personal information.

Human emotion recognition is one of the promising applications in the era of artificial super intelligence. Thus far, facial expression traits are considered to be the most widely used information cues for realizing automated emotion recognition. This paper proposes a novel facial expression recognition (FER) method that works well for recognizing emotion from image sequences. To this end, we develop the so-called weighted soft voting classification (WSVC) algorithm. In the proposed WSVC, a number of classifiers are first constructed using different and multiple feature representations. In next, multiple classifiers are used for generating the recognition result (namely, soft voting) of each face image within a face sequence, yielding multiple soft voting outputs. Finally, these soft voting outputs are combined through using a weighted combination to decide the emotion class (e.g., anger) of a given face sequence. The weights for combination are effectively determined by measuring the quality of each face image, namely "peak expression intensity" and "frontal-pose degree". To test the proposed WSVC, CK+ FER database was used to perform extensive and comparative experimentations. The feasibility of our WSVC algorithm has been successfully demonstrated by comparing recently developed FER algorithms.

In recent years, research about Natural User Interface (NUI) has become focused because NUI system can give natural feelings for users in virtual reality. Most important thing in NUI system is how to communicate with the computer system. There are many things to interact with users such as speech, hand gestures, body actions. Among them, hand gesture is suitable for the purpose of NUI because people often use a relatively high frequency in daily life and hand gesture have meaning only by itself. This hand gestures called multi-cultural hand gesture and we proposed the method to recognize this kind of hand gestures. Proposed method is composed of Histogram of Oriented Gradients (HOG) used for hand shape recognition and Histogram of Oriented Displacements (HOD) used for hand center point trajectory recognition.

Sensors in a smartphone can be used to measure various physical quantities. In this paper, we propose an object localization scheme in a three dimenstional using a smart phone. The proposed scheme estimates the location of an object by observing it from several different points. The direction to the target object and the locations of the observation points are collected at each observation point using the location sensor and the orientation sensor in the smartphone. Based on these observations, the proposed scheme derives three dimensional line of sight vectors and estimates the location of the target object that minimizes the estimation error. We implemented the proposed scheme on an Android smartphone and tested its performance by estimating the height of a building and characteristics of the proposed approach.

Recently, people's attention and worries about fine particulate matter have been increasing. Due to the construction and maintenance costs, there are insufficient air quality monitoring stations. As a result, people have limited information about the concentration of fine particulate matter, depending on the location. Studies have been undertaken to estimate the fine particle concentrations in areas without a measurement station. Yet there are limitations in that the estimate cannot take account of other factors that affect the concentration of fine particle. In order to solve these problems, we propose a framework for estimating the concentration of fine particulate matter of a specific area using meteorological data and traffic data. Since there are more grids without a monitor station than grids with a monitor station, we used a domain adversarial neural network based on the domain adaptation method. The features extracted from meteorological data and traffic data are learned in the network, and the air quality index of the corresponding area is then predicted by the generated model. Experimental results demonstrate that the proposed method performs better as the number of source data increases than the method using conditional random fields.

Because of the open data in public domains, it is easy to collect the public data. However, people find difficulty in visualizing the data in material that they actually want. Especially, A map is a difficult material to do the visualization work, without using the specific tools and learning. Therefore, in this paper, we proposed an interface tool for map data visualization that user can easily visualize various national statistics data on the map. We designed the interface by classifying the properties of the map systematically, focusing on the completion and convenience for making the map. After that, we developed a web-based application using D3.js. After user evaluation, we found that our application can visualize the map more quickly and completely than any other web interfaces for map data visualization. Users also found a higher satisfaction in operating convenience.

Recently, augmented reality technology has received much attention in many fields. This paper presents an augmented reality system using Rubiks' Cube which can augment various 3D objects depending on patterns of a Rubiks' cube. The system first detects a cube from an image using partitional clustering and strongly connected graph. Thereafter, the system detects the top side of the cube and finds a proper pattern to determine which object should be augmented. An object corresponding to the pattern is finally augmented according to the camera viewpoint. Experimental results show that the proposed system successfully augments various virtual objects in real time.

In this paper, we propose a method for evaluating the uncomfortable shaking in the video. When you shoot a video using a handheld device, such as a smartphone, most of the video contains unwanted shake. Most of these fluctuations are caused by hand tremors that occurred during shooting, and many methods for correcting them automatically have been proposed. It is necessary to evaluate the shake correction performance in order to compare the proposed shake correction methods. However, since there is no standardized performance evaluation method, a correction performance evaluation method is proposed for each shake correction method. Therefore, it is difficult to make objective comparison of shake correction method. In this paper, we propose a method for objectively evaluating video shake. Automatically analyze the video to find out how much tremors are included in the video and how much the tremors are concentrated at a specific time. In order to measure the shaking index, we proposed jitter modeling. We applied the algorithm implemented by Optical Flow to the real video to automatically measure shaking frequency. Finally, we analyzed how the shaking indices appeared after applying three different image stabilization methods to nine sample videos.

We propose a frequency estimation algorithm of human movements using Kinect. We collect the 3D coordinates of the joints of a human body and then obtain the frequency-domain description of the movements using the discrete Fourier transform (DFT). By choosing the frequency with the biggest magnitude in the selected frequencies of each of human's joint, we obtain the major beat of the human movements. Experimental results show that the proposed algorithm accurately estimates the frequency of human movements. We expect that the proposed algorithm would be applied to many AR and VR applications as a preprocessing.

'Semantic gap' has been a year-old problem in automatic video summarization, which refers to the gap between semantics implied in video summarization algorithms and what people actually infer from watching videos. Using the external EEG bio-feedback obtained from video watchers as a solution of this semantic gap problem has several another issues: First, how to define and measure noises against ERP waveforms as signals. Second, whether individual differences among subjects in terms of noise and SNR for conventional ERP studies using still images captured from videos are the same with those differently conceptualized and measured from videos. Third, whether individual differences of subjects by noise and SNR levels help to detect topic-irrelevant shots as signals which are not matched with subject's own semantic topical expectations (mis-match negativity at around 400m after stimulus on-sets). The result of repeated measures ANOVA test clearly shows a 2-way interaction effect between topic-relevance and noise level, implying that subjects of low noise level for video watching session are sensitive to topic-irrelevant visual shots, while showing another 3-way interaction among topic-relevance, noise and SNR levels, implying that subjects of high noise level are sensitive to topic-irrelevant visual shots only if they are of low SNR level.

It is essential to collect and analyze target information around the vehicle for autonomous driving of the vehicle. Based on the analysis, environmental information such as location and direction should be analyzed in real time to control the vehicle. In particular, obstruction or cutting of objects in the image must be handled to provide accurate information about the vehicle environment and to facilitate safe operation. In this paper, we propose a method to simultaneously generate 2D and 3D bounding box proposals using LiDAR Edge generated by filtering LiDAR sensor information. We classify the classes of each proposal by connecting them with Region-based Fully-Covolutional Networks (R-FCN), which is an object classifier based on Deep Learning, which uses two-dimensional images as inputs. Each 3D box is rearranged by using the class label and the subcategory information of each class to finally complete the 3D bounding box corresponding to the object. Because 3D bounding boxes are created in 3D space, object information such as space coordinates and object size can be obtained at once, and 2D bounding boxes associated with 3D boxes do not have problems such as occlusion.

Since the mitigation of fear of crime significantly enhances the consumptions in a city, studies focusing on urban safety analysis have received much attention as means of revitalizing the local economy. In addition, with the development of computer vision and machine learning technologies, efficient and automated analysis methods have been developed. Previous studies have used global features to predict the safety of cities, yet this method has limited ability in accurately predicting abstract information such as safety assessments. Therefore we used a Convolutional Context Neural Network (CCNN) that considered "context" as a decision criterion to accurately predict safety of cities. CCNN model is constructed by combining a stacked auto encoder with a fully connected network to find the context and use it in the CNN model to predict the score. We analyzed the RMSE and correlation of SVR, Alexnet, and Sharing models to compare with the performance of CCNN model. Our results indicate that our model has much better RMSE and Pearson/Spearman correlation coefficient.

Optical coherence tomography (OCT) is the most common medical imaging device with the ability to image the retina in eyes at micrometer resolution and to visualize the pathological indicators of many retinal diseases such as Age-Related Macular Degeneration (AMD) and diabetic retinopathy. Accordingly, there have been research activities to analyze and process OCT images to identify those indicators and make medical decisions based on the findings. In this paper, we use a deep convolutional neural network for analysis of OCT volume data to distinguish AMD from normal patients. We propose a novel approach where images in each OCT volume are grouped together into sub-volumes and the network is trained by those sub-volumes instead of individual images. We conducted an experiment using public data set to evaluate the performance of the proposed approach. The experiment confirmed performance improvement of our approach over the traditional image-by-image training approach.

Sounds based machine fault diagnosis recovers all the studies that aim to detect automatically faults or damages on machines using the sounds emitted by these machines. Conventional methods that use mathematical models have been found inaccurate because of the complexity of the industry machinery systems and the obvious existence of nonlinear factors such as noises. Therefore, any fault diagnosis issue can be treated as a pattern recognition problem. We present here an automatic fault diagnosis system of hand drills using discrete wavelet transform (DWT) and pattern recognition techniques such as principal component analysis (PCA) and artificial neural networks (ANN). The diagnosis system consists of three steps. Because of the presence of many noisy patterns in our signals, we first conduct a filtering analysis based on DWT. Second, the wavelet coefficients of the filtered signals are extracted as our features for the pattern recognition part. Third, PCA is performed over the wavelet coefficients in order to reduce the dimensionality of the feature vectors. Finally, the very first principal components are used as the inputs of an ANN based classifier to detect the wear on the drills. The results show that the proposed DWT-PCA-ANN method can be used for the sounds based automated diagnosis system.

Recently, vector map data is developed and used in many domains widely. In the most cases, vector map data contains confidential information which must be kept away from unauthorized users. Moreover, the production process of vector maps is considerably complex and consumes a lot of money and human resources. Therefore, the secured storage and transmission are necessary to prevent the illegal copying and distribution from hacker. This paper presents a selective encryption algorithm using geometric objects in frequency domain for vector map data. In the proposed algorithm, polyline and polygon data in vector map is the target of the selective encryption process. Experimental results verified that proposed algorithm is effectively and adaptive the requirements of security.

A quantum computer, based on quantum mechanics, is a paradigm of information processing that can show remarkable possibilities of exponentially improved information processing. This paradigm can be solved in a short time by calculating factoring problem and discrete logarithm problem that are typically used in public key cryptosystems such as RSA(Rivest-Shamir-Adleman) and ECC(Elliptic Curve Cryptography). In 2013, Lei et al. proposed a secure NTRU-based key distribution protocol for quantum computing. However, Lei et al. protocol was vulnerable to man-in-the-middle attacks. In this paper, we propose a NTRU(N-the truncated polynomial ring) key distribution protocol with mutual authentication only using NTRU convolution multiplication operation in order to maintain the security for quantum computing. The proposed protocol is resistant to quantum computing attacks. It is also provided a secure key distribution from various attacks such as man-in-the middle attack and replay attack.

In off-line phase of the preliminary Cut-off indoor positioning scheme, which is one of the indoor positioning scheme using the fingerprint, relative ranks of peak RSSIs received from beacons at each reference point are stored in the fingerprint map. In some reference points, signals of multiple beacons may be received. In this case, the relative ranks may be different when constructing fingerprint and when receiving signals in real-time. To solve this problem, we propose a method to utilize only up to five beacons with high ranking when constructing a fingerprint and when receiving signals in real-time and comparing them with stored information of a fingerprint. Experiments were conducted on the estimation probabilities and the average error when using this method. Those are compared with the previous methods. Experimental results show that the estimation probabilities and the average error are improved by removing only the remaining five beacons at each reference point of the fingerprint.

To achieve software quality and human-centred design for electronic ship navigation called e-navigation, an international guideline of software quality assurance and human-centred design was approved in 2015. Usability is a common goal of both software quality assurance and human-centred design as developing e-navigation system and software developments. Therefore, research is needed to evaluate the usability of e-navigation systems and software such as metrics that can use usability testing. This paper derives effectiveness metrics for e-Navigation usability testing based on international standards. The research method is to analyses and compares the effectiveness measurement and metrics in ISO 9241-11 for human-centered design and ISO/IEC 25022 and 25023 for software quality to find out measurements and metrics being defined commonly. The derived metrics are applied to Electronic Chart Display and Information System as a case study based on performance standard.

Artificial intelligence-based services are expanding into a new industrial revolution. There is artificial intelligence technology applied in real life due to the development of big data and deep learning related technology. And data analysis and intelligent assistant services that integrate information from various fields have also been commercialized. Chatbot with interactive artificial intelligence provide shopping, news or information. Chatbot service, which has begun to be adopted by some public institutions, is now just a first step in the steps. This study summarizes the services and technical analysis of chatbot. and the direction of public administration service chatbot was presented.

Sensor networks are used in various fields such as marine, defense, and smart home etc. Among the components of the sensor network, the sensor node collects sensor data, as one of the representative sensor network roles, and the sensor node makes a greate influence on the overall performance of the sensor network. Therefore, how to design the sensor node is an important issue in the sensor network field. However, the research on the sensor network architecture suitable for the sensor network installation environment has been made more important than the research on how to configure the sensor node. In this paper, we propose to identify elements to be considered for designing a sensor node that makes a large influence on the performance of the sensor network, and to easily implement the sensor node through the state transition model based on these elements.

Pub/Sub (Publish/Subscribe) paradigm is a simple and easy to use model for interconnecting applications in a distributed environment. In general, subscribers register their interests in a topic or a pattern of events and then asynchronously receive events matching their interest, regardless of the events' publisher. In order to build a low latency lightweight pub/sub system for Internet of Things (IoT) services, we propose a GQSFPS (Group Quorum System-based Fog Pub/Sub) system that is a core component in the event-driven service oriented architecture framework for IoT services. The GQSFPS organizes multiple installed pub/sub brokers in the fog servers into a group quorum based P2P (peer-to-peer) topology for the efficient searching and the low latency accessing of events. Therefore, the events of IoT are cached on the basis of group quorum, and the delay-sensitive IoT applications of edge devices can effectively access the cached events from group quorum fog servers in low latency. The performance of the proposed GQSFPS is evaluated through an analytical model, and is compared to the GQPS (grid quorum-based pud/sub system).

As the Internet of Things evolves, various IoT services are provided. IFTTT is an abbreviation for If This Then That and refers to a service that links different web-based services. This paper proposes a system that generates and manages rules that combine the possibility of IFTTT service and the real-time event processing according to the concept of IoT service. Conventional database-based data processing methods are burdened to process a lot of data of IoT devices coming in real-time. The IoT device's data can be classified into formal data such as the amount of power, temperature value and position information, and informal data such as video or image data. Thus, this system classifies the data stream of IoT devices coming in real-time using the CEP engine Esper into a file signature table, classifies the formal/informal data, and shows the condition of the device data defined by the user and the service to be provided by applying the service.

As rate in an elderly population and expanding their range of activity rapidly increase, the demographics of the elderly population on a use of transportation also rise. Elderly pedestrians often find it difficult to react promptly to the traffic accidents as they are less perceptive of the dangers present under the situation. More than half of the elderly traffic accidents are elderly pedestrian accidents in road. Therefore, we design and implement smart safety hat for safety of elderly pedestrian. The smart safety hat binds stripe-shaped LED around a hat in order for a driver to perceive pedestrian easy and quickly. Features of smart safety hat include controlling the number of LEDs by using a light sensor and warning through vibration using a sound sensor. Also, we used Bluetooth to communicate with the smartphone to enable user customization of the light and numbers of LEDs.

In this paper, we analyses heart beat data to identify subjects stress state (binary) using heart rate variability (HRV) features extracted from heart beat data of the subjects and implement supervised machine learning techniques to create the mental stress classifier. There are four steps need to be done: data acquisition, data processing (HRV analysis), features selection, and machine learning, before doing performance measurement. There are 56 features generated from the HRV Analysis module with several of them are selected (using own algorithm) after computing the Pearson Correlation Matrix (p-values). The results of the list of selected features compared with all features data are compared by its model error after training using several machine learning techniques: support vector machine, decision tree, and discriminant analysis. SVM model and decision tree model with using selected features shows close results compared to using all recording by only 1% difference. Meanwhile, the discriminant analysis differs about 5%. All the machine learning method used in this works have 90% maximum average accuracy.

Artificial neural networks (ANNs) play an important role in the fields of function approximation, prediction, and classification. ANN performance is critically dependent on the input parameters, including the number of neurons in each layer, and the optimal values of weights and biases assigned to each neuron. In this study, we apply the particle swarm optimization method, a popular optimization algorithm for determining the optimal values of weights and biases for every neuron in different layers of the ANN. Several regression models, including general linear regression, Fourier regression, smoothing spline, and polynomial regression, are conducted to evaluate the proposed method's prediction power compared to multiple linear regression (MLR) methods. In addition, residual analysis is conducted to evaluate the optimized ANN accuracy for both training and test datasets. The experimental results demonstrate that the proposed method can effectively determine optimal values for neuron weights and biases, and high accuracy results are obtained for prediction applications. Evaluations of the proposed method reveal that it can be used for prediction and estimation purposes, with a high accuracy ratio, and the designed model provides a reliable technique for optimization. The simulation results show that the optimized ANN exhibits superior performance to MLR for prediction purposes.

Human activity recognition has been studied using various sensors and algorithms. Human activity recognition can be divided into sensor based and vision based on the method. In this paper, we proposed an activity recognition system using acceleration sensor and gyroscope sensor in smartphone among sensor based methods. We used Deep Belief Network (DBN), which is one of the most popular deep learning methods, to improve an accuracy of human activity recognition. DBN uses the entire input set as a common input. However, because of the characteristics of different time window depending on the type of human activity, the RBMs, which is a component of DBN, are configured hierarchically by combining them from different time windows. As a result of applying to real data, The proposed human activity recognition system showed stable precision.

Recently, interest that intra-abdominal pressure has been increased as change of pathophysiology to critical patients. The intra-abdominal pressure is measured by cystometry what can be available for non-inclusively. However, conventional methods have some problems such as low SNR, weakness of environment temperature, and unsuitable size of sensor. In this paper, a new subminiature pressure sensor module and sensing system are proposed using a sensor of semiconductor type and FPCB. The module is more stable, flexible, and smaller than the conventional catheter. The performance of the developed module is evaluated by various quantitative analysis indexes. The proposed sensor has the high sensitivity and suitable size for measurement of cystometry more than the conventional method. In order to prove efficiency between conventional and proposed method, proposed method compared for sensitivity, fixable, and size. The proposed method will be help measurement of intra-abdominal pressure of patients due to high accuracy and comfortableness.

Steady exercise or walking exercise is helpful for the treatment of chronic diseases or cancers. In this paper, I presented a smart poster to enable the patients to exercise while moving between the smart posters, dynamically, in order to provide better exercise effect to them. It can be a new form of exercise prescription that combines exercise with walking using smart posters. The personalized exercise prescription is downloaded from the management server in real time when the patient approaches, and induces the patient's exercise and walking. In addition, the smart poster helps patient to move to other posters in order to induce more walking exercise. To achieve this, I proposed a transfer protocol that autonomously exchanges session information between smart posters in this paper. Moreover, the smart poster based on Raspberry was implemented to verify validity of this protocol, and an experiment was conducted to measure the request and response time between smart posters in the implemented environment. In the experiment, when the other poster sent the message requesting the exercise session 100 times and received the response message, the 95 percentage of received messages had the response time within 0.05 seconds.

In a trawl-net simulation, it is very important to process the physical phenomenons resulting from real collisions between a net and fishes. However, because it is very difficult to reconstruct the surface with mass points, many researchers have generally detect the collision using an approximation model employing a sphere, a cube or a cylinder. These approaches occur often result in inaccurate movements of a fish due to the difference between a real-net and a designed-net. So, many systems have manually adjusted a net surface based on actual measurements of mass points. These methods are very inefficient because it needs much times in an adjustment and also causes more incorrect inputs according to a rapid increment in the number of points. Therefore, in this paper, we propose a reconstruction method that it semi-automatically reconstructed trawl-net surfaces using the equation of motion at each mass point in a mass-spring model. To get an easy start in a beginning step of the spread, it enables users to get interactive adjustment on each mass point. We had designed a trawl-net model using geometrical structures of trawl-net and then automatically reconstructed the trawl-net surface using scale-space meshing techniques. Last, we improve the accuracy of reconstructed result by correction user interaction.

In recent times, IaaS (Infrastructure as a Services) have been rapidly evolving to allow developers to easily and efficiently access work in the server and network areas for development of a web of App based on cloud computing. In this study, we developed the collaborative image editing tool App based on Cloud-computing, by adopting AWS of representative company that develops IaaS. First, it is crucial to understand various situation conditions for representative infrastructure services: AWS, Azure and Google (GCP). This may have the effect of reducing manpower and development time, but as each company has different policy and technical support, we need a new study every time the environment changes of infrastructure services. We tried to develop a hybrid-App so that users with various devices can collaborate work each other by utilizing the infrastructure service AWS through the process of developing the image editing authoring tool based on the cloud computing. The future studies should continue about compatibility issues and support issues in order to minimize the problems of overseas infrastructure services, but we think that domestic cloud computing policies and developments should be urgently considered.

One of the important factors in game design is difficulty adjustment. An appropriate level of difficulty makes users have a sense of challenge and interest. However, the adjustment of difficulty takes a lot of time and effort, because of its ambiguity. To solve the problem, we propose a difficulty control method using a fuzzy theory. In this paper, a simple demonstration is exemplified to verify the effectiveness of our method. Experimental results show that the difficulty of the game changes according to the user's skill.

The trailer is an advertising process that requires systematic fabrication in the planning process. However, the role of sufficient consideration of the specialty of the merchandise, and the study of the quality of the merchandise, can serve as an effective advertisement. Therefore, a trailer from a marketing standpoint should effectively express the components that need to be delivered to potential audiences within a brief running time. Want to see if the rhythm of the trailers can be tangible, and how it affects receivers.