This paper presents a unified framework for joint Convolutional Neural Network (CNN) based vehicle detection by leveraging multi-spectral image pairs. With the observation that under challenging environments such as night vision and limited light source, vehicle detection in a single color image can be more tractable by using additional far-infrared (FIR) image, we design joint CNN architecture for both RGB and FIR image pairs. We assume that a score map from joint CNN applied to overall image can be considered as confidence of vehicle existence. To deal with various scale ratios of vehicle candidates, multi-scale images are first generated scaling an image according to possible scale ratio of vehicles. The vehicle candidates are then detected on local maximal on each score maps. The generation of overlapped candidates is prevented with non-maximal suppression on multi-scale score maps. The experimental results show that our framework have superior performance than conventional methods with a joint framework of multi-spectral image pairs reducing false positive generated by conventional vehicle detection framework using only single color image.

The people who complain of pain due to voice causes such as vocal cord nodules is increasing year by year. If the voice is changed, it is possible to give to colleagues discomfort or inconvenience during conversation. In this paper, we propose a way to reduce discomfort by improving the husky voice during the conversation. A VBF (voice boosting filter) is firstly designed to improve the husky voices. This filter may further emphasize the formant frequency components than the frequency components around the formant frequency, because the value is relatively greater than the other frequency. And a fixed-point type DSP chipset, TMS320F2812 is applied to the system, the operating frequency is 150MHz. The system was implemented as a compact for use as a portable, its size is $2.5cm{\times}10cm$. Through the test using three husky voices with some type of statement, it was satisfactory in processing speed and sound quality improvement.

This paper propose a framework for human body parts in RGB-D image. We conduct tasks of obtaining person area, finding candidate areas and local detection in order to detect hand, foot and head which have features of long accumulative geodesic distance. A person area is obtained with background subtraction and noise removal by using depth image which is robust to illumination change. Finding candidate areas performs construction of graph model which allows us to measure accumulative geodesic distance for the candidates. Instead of raw depth map, our approach constructs graph model with segmented regions by quadtree structure to improve searching time for the candidates. Local detection uses HOG based SVM for each parts, and head is detected for the first time. To minimize false detections for hand and foot parts, the candidates are classified with upper or lower body using the head position and properties of geodesic distance. Then, detect hand and foot with the local detectors. We evaluate our algorithm with datasets collected Kinect v2 sensor, and our approach shows good performance for head, hand and foot detection.

This paper presents a mobile color-compensating application using a RGB color compensation algorithm. The proposed application enables color vision deficiencies to differentiate images which cannot be distinguished by them. The suggested mobile application has two functions: converting images taken by camera and converting contents in the device screen. The proposed application is computationally inexpensive, since it does not require color space transformation. Simulation results show that the proposed application enables color vision deficiencies to receive information expressed by colors such as subway line maps.

Recently, gender classification has attracted a great deal of attention in the field of video surveillance system. It can be useful in many applications such as detecting crimes for women and business intelligence. In this paper, we proposed a method which can detect pedestrians from CCTV video and classify the gender of the detected objects. So far, many algorithms have been proposed to classify people according the their gender. This paper presents a gender classification using convolutional neural network. The detection phase is performed by AdaBoost algorithm based on Haar-like features and LBP features. Classifier and detector is trained with data-sets generated form CCTV images. The experimental results of the proposed method is male matching rate of 89.9% and the results shows 90.7% of female videos. As results of simulations, it is shown that the proposed gender classification is better than conventional classification algorithm.

In this paper, we propose a two-scale fusion method for infrared and visible images using saliency and variance. The images are separated into two scales respectively: a base layer of low frequency component and a detailed layer of high frequency component. Then, these are synthesized using weight. The saliencies and the variances of the images are used as the fusion weights for the two-scale images. The proposed method is tested on several image pairs, and its performance is evaluated quantitatively by using objective fusion metrics.

In this paper, we propose an automatic hierarchical organ segmentation method on abdominal CT images. First, similar atlases are selected using bone-based similarity registration and similarity of liver, kidney, and pancreas area. Second, each abdominal organ is roughly segmented using image-based similarity registration and intensity-based locally weighted voting. Finally, the segmented abdominal organ is refined using mask-based affine registration and intensity-based locally weighted voting. Especially, gallbladder and pancreas are hierarchically refined using location information of neighbor organs such as liver, left kidney and spleen. Our method was tested on a dataset of 12 portal-venous phase CT data. The average DSC of total organs was $90.47{\pm}1.70%$. Our method can be used for patient-specific abdominal organ segmentation for rehearsal of laparoscopic surgery.

The problem of food deficiency is a major discouragement to many low-income developing countries. Most of these countries experience constant danger of hunger, malnutrition and diseases as they are unable to maintain their food supplies mainly due to lack of arable lands and modern crop, livestock and fishery production technologies. In addition, the pollutants resulting from the secondary industries are becoming another serious issue in their food problems. The pollutants mixed in the sands blowing from the mainland China and the toxic waters flowing in the farm land form the industrialized zones are some of the examples. The Vertical Farm, or Plant Factory, proposed in this study could be the best alternative food production system for them. Vertical farm is an efficient food production system that yields relatively a large volume of food materials without environmental risks. The system does not require a large open space and manpower and can minimize the possibility of infiltration of pollutants. This research describes a basic model of the system focusing on determining the optimal sites for it based on the meteorological data concentrating on the atmospheric pollutants. The types and volume of pollutants are analyzed and identified through the big data obtained, followed by visualization of analysis results and their comparisons for better understanding.

In recent, the advancement of wearable devices and wireless body area networking technologies motivate researchers to pay attention to remote healthcare system for monitoring patients health and disease progression effectively. However, in order to implement a practical remote healthcare system, we must consider the security and privacy of patient's personal health information transmitted to healthcare servers through the network. Hence, in this paper, we propose a secure health data transmission protocol in remote healthcare monitoring system to protect patient's health information and prevent privacy from eavesdropping on the network. To achieve our security goals, we design an efficient secure protocol based on the identity-based cryptography with key evolution technique, and then confirm the superiority and the efficiency of the proposed protocol as compared with the existing protocol of Yang et al.

In layered communication architecture, each layer is designed to service its own functions to higher layer while getting serviced by lower layer. Usually layered architectures are not optimized in a total view of whole services and functions. So cross layer design pursues performance enhancements by optimizing in various ways. In LTE, MAC layer uses HARQ mechanism and RLC layer uses ARQ mechanism for retransmission. According to the 3GPP 36.331 specification, two layers' cooperation may not happen in an optimized way. This paper suggests an adaptive MAC layer approach which RLC layer's function might be initiated in MAC layer in advance to utilize MAC layer's idling wasting time for RLC layer's next decision. This adaptive ARQ method in MAC layer speeds up the next retransmission and reduces the overall transmission time. Emulation shows the improved performance in total retransmission time and retransmission success ratio. In wireless shadow area, the retransmission occurs frequently. Our approach has strong points in this poor wireless condition.

The purpose of this study is to validate the validity of the methodology for analyzing the space with complex characteristics and to evaluate the existing spatial structure analysis theory. Seven example models are designed and analyzed data of spatial syntax analysis and visibility graph analysis. And analyzed the agent-based model using two analytical methods: the adjacent space and the whole spatial connection. The results of this study are as follows. Based on the analysis of the agent - based model for perfectly freewalking, the validity of the method is verified in terms of predictive ability and effectiveness. Agent-based models can be simulated considering various variables, so realistic predictions will be possible and a new biography of complex systems can be met.

Web presence is one of the key issues for extensive deployment of Internet-of-Things (IoT). An obstacle to overcome for Web presence is relatively low computing power of IoT devices. In this paper, we present implementation of an IoT platform based on Constrained Application Protocol (CoAP) which is a web transfer protocol proposed by Internet Engineering Task Force (IETF) for the low performance IoT devices such as Wireless Sensor Network (WSN) nodes and micro-controllers. To qualify the performance of CoAP-based IoT system for such an application as smart grid, we designed a test platform consisting of Raspberry Pi2, Kmote WSN node and a desktop PC. Using open source softwares, CoAP was implemented on top of the platform. Leveraging the GET command defined at CoAP specification, performance of the system was measured in terms of round-trip time (RTT) from web application to the Kmote sensor node. To investigate abnormal cases among the test results, hop-by-hop delays were measured to analyze resulting data. The average response time of CoAP-based communication except the abnormal data was reduced by 23% smaller than the previous research result.

Modern industry requires people with the ability to create and improve their knowledge. Most educational institutions have used mentoring activities which can train those people. Then each of the institutions has different directions because they work in their own fields. For that reason, many people which want to make new mentoring program have no choice but to go through many trials and errors. This paper deals with the design method of basic process model for mentoring based on DEVS in order that people plan a new mentoring business or make a new mentoring group easily. Our proposed model is expected to be used as a guide to establish procedures for mentoring systems.

The $360^{\circ}$ Virtual Reality (VR) live-action movies are filmed by attaching multiple cameras to a rig to shoot the images omni-directionally. Especially, for a live-action film that requires a variety of scenes, the director of photography and his staff usually have to operate the rigged cameras directly all around the scene and edit the footage during the post-production stage so that the entire process can incur much time and high cost. However, it will also be possible to acquire high-quality omni-directional images with fewer staff if the camera rig(s) can be controlled remotely to allow more flexible camera walking. Thus, a $360^{\circ}$ VR filming system with remote-controlled camera rig has been proposed in this study. The movie producers will be able to produce the movies that provide greater immersion with this system.