Earticle

Harvested de-astringent persimmon ‘Fuyu’ were treated with various lighting sources under low (3℃) and high (22℃ ) temperatures. The weight loss rate of fruits was lower in those with Red LED than Fluorescence and Blue LED under both temperature conditions. Hardness and soluble solid content of fruits were higher in those with 3℃ / Blue LED or mixed LED (Blue+Red LEDs). Beta-carotene and lycopene content of fruit peel were higher in those with 3℃ than 22℃ and with Red LED or light sources with mixed red wavelength under both temperatures. When the fruits treated with light and temperature were stored for 4 days under 3℃ / dark condition, the hardness of the fruits did not significant difference among the treatments. Taken together all the results, it would be best to treat it light sources mixed red wavelength under 3℃.

The apparent soil resistivity is used for estimating multilayer soil parameters, such as, layer’s depth and soil resistivity. The soil parameters are estimated by continuously revising those parameters until the error between the measured and calculated apparent soil resistivity reaches to allowable level. The equation for calculating the apparent soil resistivity is complicated and time consumed, because it is composed of an infinite integral which includes a zero order Bessel’s function of the first kind. In this paper, a fast algorithm for calculating the apparent soil resistivity of horizontal multilayer earth structure is proposed using exponential sampling method.

This paper discusses the main requirements in today's full-color holograms illumination and evaluates the last generation of LEDs, the actual best light source to render properly the colors of the holograms and in particular those recorded with red 660 nm, green 532 nm and blue 440 nm lasers. This paper presents also the first prototype of lamp designed especially for this kind of holograms.

Neural networks have been reborn as a Deep Learning thanks to big data, improved processor, and some modification of training methods. Neural networks used to initialize weights in a stupid way, and to choose wrong type activation functions of non-linearity. Weight initialization contributes as a significant factor on the final quality of a network as well as its convergence rate. This paper discusses different approaches to weight initialization. MNIST dataset is used for experiments for comparing their results to find out the best technique that can be employed to achieve higher accuracy in relatively lower duration.

As social network service platforms grow and one-person media market expands, people upload their own photos and/or videos through multiple open platforms. However, it can be illegal to upload the digital contents containing the faces of others on the public sites without their permission. Therefore, many people are spending much time and effort in editing such digital contents so that the faces of others should not be exposed to the public. In this paper, we propose an automatic face hiding system called ‘autoblur’, which detects all the unregistered faces and mosaic them automatically. The system has been implemented using the GitHub MIT open-source ‘Face Recognition’ which is based on deep learning technology. In this system, two dozens of face images of the user are taken from different angles to register his/her own face. Once the face of the user is learned and registered, the system detects all the other faces for the given photo or video and then blurs them out. Our experiments show that it produces quick and correct results for the sample photos.

Since the third artificial intelligence boom was triggered by deep learning, it has been 10 years. It is time to analyze and discuss the research trends of deep learning for the stable development of AI. In this regard, this study systematically analyzes the trends of research on deep learning over the past 10 years. We collected research literature on deep learning and performed LDA based topic modeling analysis. We analyzed trends by topic over 10 years. We have also identified differences among the major research countries, China, the United States, South Korea, and United Kingdom. The results of this study will provide insights into research direction on deep learning in the future, and provide implications for the stable development strategy of deep learning.

In this paper, we propose an efficient method for AR toolkit Vuforia. In order to increase the scan rate when using the 3D object scanner, the scan rate parameters need to be analyzed in terms of the angle and distance. In addition, in order to increase the tracking rate when tracking an object, the tracking rate has to be evaluated according to the position, complexity, and contrast of the object. To this end, we have defined the difference of scan rate according to angle and distance between camera and object when using object scanner and the recognition time according to object's position, complexity and contrast when tracking object.

When one enters a space, perceives the material geometry of that space. Walking inside buildings or across the city is generating a geometry of moving bodies that fills the space. These two geometries coexist: a static geometry of the space and an invisible one of the moving bodies. The space that we actually experience, whether interior or exterior, is a continuous network of voids. Individuals’ movement will fill the network of voids that we understand as “the city”. Our environment of voids and borders is organized by the means of architecture and urbanism. The geometry generated by motion affects both the limits and the voids, thus space can be defined by the tandem of the moving bodies and their environment. We propose in this study a mean of investigating users’ movement and thus understanding the qualities of space while introducing the concept of space scores as analytical maps and design tools.

Nowadays, as the availability of tiny, low-cost microcomputer increases at a high level, mobile robots are experiencing remarkable enhancements in hardware design, software performance, and connectivity advancements. In order to control Turtlebot 2, several algorithms have been developed using the Robot Operating System(ROS). However, ROS requires to be run on a high-cost computer which increases the hardware cost and the power consumption to the robot. Therefore, design an algorithm based on low-cost hardware is the most innovative way to reduce the unnecessary costs of the hardware, to increase the performance, and to decrease the power consumed by the computer on the robot. In this paper, we present a path-finding algorithm for TurtleBot 2 based on low-cost hardware. We implemented the algorithm using Raspberry pi, Windows 10 IoT core, and RPLIDAR A2. Firstly, we used Raspberry pi as the alternative to the computer employed to handle ROS and to control the robot. Raspberry pi has the advantages of reducing the hardware cost and the energy consumed by the computer on the robot. Secondly, using RPLIDAR A2 and Windows 10 IoT core which is running on Raspberry pi, we implemented the path-finding algorithm which allows TurtleBot 2 to navigate from the starting point to the destination using the map of the area. In addition, we used C# and Universal Windows Platform to implement the proposed algorithm.

As AI technology is recently introduced into various fields, it is being applied to the fashion field. This paper proposes a system for recommending cody clothes suitable for a user's selected clothes. The proposed system consists of user app, cody recommendation module, and server interworking of each module and managing database data. Cody recommendation system classifies clothing images into 80 categories composed of feature combinations, selects multiple representative reference images for each category, and selects 3 full body cordy images for each representative reference image. Cody images of the representative reference image were determined by analyzing the user's preference using Google survey app. The proposed algorithm classifies categories the clothing image selected by the user into a category, recognizes the most similar image among the classification category reference images, and transmits the linked cody images to the user's app. The proposed system uses the ResNet-50 model to categorize the input image and measures similarity using ORB and HOG features to select a reference image in the category. We test the proposed algorithm in the Android app, and the result shows that the recommended system runs well.

Cluster computing are more utilized is the fragment of day by day indoctrination life. Imagining variations after incorporating server processors to client server establishment or cluster computing insurances after system, enrolling, Bank process, utility imagining also automatic considering along with an impelled association step. The rapid variations just before the cloud, takes certain fundamental concerns intended for achievement of information structures, records exchanges and material safety. Starting safety problem cloud applications take couple of huge issues for the reasonability of bunch assuming the equipment, gadgets, Encryption programming, firewalls, verification software’s the cloud applications use resources, handling memory and limit are definitely not physically available at customer registering. This will help the customer to affliction recovery. Distributed computing is mostly used worldwide in which the distinctive errands are consigned to a blend of affiliations, programming and organizations that can be gotten over the framework. The cloud applications handle consumer data loads to various appropriated areas where utility, availability and portability are feasible. Our proposed algorithm will reduce the delay factor to allocate the resources by random selection mapping technique among the Cluster based Network which will in future helpful for grid computing performance analysis.

Wind speed data constitute important weather information for aircrafts flying at low altitudes, such as drones. Currently, the accuracy of low altitude wind predictions is much lower than that of high-altitude wind predictions. Deep neural networks are proposed in this study as a method to improve wind speed forecast information. Deep neural networks mimic the learning process of the interactions among neurons in the brain, and it is used in various fields, such as recognition of image, sound, and texts, image and natural language processing, and pattern recognition in time-series. In this study, the deep neural network model is constructed using the wind prediction values generated by the numerical model as an input to improve the wind speed forecasts. Using the ground wind speed forecast data collected at the Boseong Meteorological Observation Tower, wind speed forecast values obtained by the numerical model are compared with those obtained by the model proposed in this study for the verification of the validity and compatibility of the proposed model.