• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.29 no.11
• /
• pp.96-101
• /
• 2015

In recent years, researches regarding optimal placement of CCTV(Closed-circuit Television) cameras via architecture modeling has been conducted. However, for analyzing surveillance coverage through actual human movement, the application of VA(Video Analytics) function of IP(Internet Protocol) cameras has not been studied. This paper compares two methods using data captured from real-world cameras and data acquired from a virtual environment. In using real cameras, we develop GUI(Graphical User Interface) to be used as a logfile which is stored hourly and daily through VA functions and to be used commercially for placement of products inside a shop. The virtual environment was constructed to emulate an real world such as the building structure and the camera with its specifications. Moreover, suitable placement of the camera is done by recognizing obstacles and the number of people counted within the camera's range of view. This research aims to solve time and economic constraints of actual installation of surveillance cameras in real-world environment and to do feasibility study of virtual environment.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.1 no.1
• /
• pp.39-49
• /
• 2009

The use of high-strength aluminum alloys is increasing in shipbuilding industry, particularly for the design and construction of war ships, littoral surface craft and combat ships, and fast passenger ships. While various welding methods are used today to fabricate aluminum ship structures, namely gas metallic arc welding (GMAW), laser welding and friction stir welding (FSW), FSW technology has been recognized to have many advantages for the construction of aluminum structures, as it is a low-cost welding process. In the present study, mechanical properties of friction stir welded aluminum alloys are examined experimentally. Tensile testing is undertaken on dog-bone type test specimen for aluminum alloys 5083 and 5383. The test specimen includes friction stir welded material between identical alloys and also dissimilar alloys, as well as unwelded (base) alloys. Mechanical properties of fusion welded aluminum alloys are also tested and compared with those of friction stir welded alloys. The insights developed from the present study are documented together with details of the test database. Part of the present study was obtained from the Ship Structure Committee project SR-1454 (Paik, 2009), jointly funded by its member agencies.

• Journal of Broadcast Engineering
• /
• v.24 no.1
• /
• pp.182-185
• /
• 2019

This paper analyzes performances of modules in implementing a programmable multi-format video decoder. The goal of the proposed platform is the high-end Full High Definition (FHD) video decoder. The proposed multi-format video decoder consists of a reconfigurable processor, dedicated bit-stream co-processor, memory controller, cache for motion compensation, and flexible hardware accelerators. The experiments suggest performance baseline of modules for the proposed architecture operating at 300 MHz clock with capability of decoding HEVC bit-streams of FHD 30 frames per second.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.4
• /
• pp.1-8
• /
• 2023

In this paper, we propose a black ice detection platform framework using Convolutional Neural Networks (CNNs). To overcome black ice problem, we introduce a real-time based early warning platform using CNN-based architecture, and furthermore, in order to enhance the accuracy of black ice detection, we apply a multi-scale dilation convolution feature fusion (MsDC-FF) technique. Then, we establish a specialized experimental platform by using a comprehensive dataset of thermal road black ice images for a training and evaluation purpose. Experimental results of a real-time black ice detection platform show the better performance of our proposed network model compared to conventional image segmentation models. Our proposed platform have achieved real-time segmentation of road black ice areas by deploying a road black ice area segmentation network on the edge device Jetson Nano devices. This approach in parallel using multi-scale dilated convolutions with different dilation rates had faster segmentation speeds due to its smaller model parameters. The proposed MsCD-FF Net(2) model had the fastest segmentation speed at 5.53 frame per second (FPS). Thereby encouraging safe driving for motorists and providing decision support for road surface management in the road traffic monitoring department.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.791-802
• /
• 2023

The manual inspection of photovoltaic (PV) panels to meet the requirements of inspection work for large-scale PV power plants is challenging. We present a hot spot detection and positioning method to detect hot spots in batches and locate their latitudes and longitudes. First, a network based on the YOLOv3 architecture was utilized to identify hot spots. The innovation is to modify the RU_1 unit in the YOLOv3 model for hot spot detection in the far field of view and add a neural network residual unit for fusion. In addition, because of the misidentification problem in the infrared images of the solar PV panels, the DeepLab v3+ model was adopted to segment the PV panels to filter out the misidentification caused by bright spots on the ground. Finally, the latitude and longitude of the hot spot are calculated according to the geometric positioning method utilizing known information such as the drone's yaw angle, shooting height, and lens field-of-view. The experimental results indicate that the hot spot recognition rate accuracy is above 98%. When keeping the drone 25 m off the ground, the hot spot positioning error is at the decimeter level.

• ETRI Journal
• /
• v.46 no.1
• /
• pp.22-34
• /
• 2024

Exposure to varied noisy environments impairs the recognition performance of artificial intelligence-based speech recognition technologies. Degraded-performance services can be utilized as limited systems that assure good performance in certain environments, but impair the general quality of speech recognition services. This study introduces an audiovisual speech recognition (AVSR) model robust to various noise settings, mimicking human dialogue recognition elements. The model converts word embeddings and log-Mel spectrograms into feature vectors for audio recognition. A dense spatial-temporal convolutional neural network model extracts features from log-Mel spectrograms, transformed for visual-based recognition. This approach exhibits improved aural and visual recognition capabilities. We assess the signal-to-noise ratio in nine synthesized noise environments, with the proposed model exhibiting lower average error rates. The error rate for the AVSR model using a three-feature multi-fusion method is 1.711%, compared to the general 3.939% rate. This model is applicable in noise-affected environments owing to its enhanced stability and recognition rate.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.10
• /
• pp.3989-4006
• /
• 2020

The intelligent agriculture monitoring is based on the perception and analysis of environmental data, which enables the monitoring of the production environment and the control of environmental regulation equipment. As the scale of the application continues to expand, a large amount of data will be generated from the perception layer and uploaded to the cloud service, which will bring challenges of insufficient bandwidth and processing capacity. A fog-based offline and real-time hybrid data analysis architecture was proposed in this paper, which combines offline and real-time analysis to enable real-time data processing on resource-constrained IoT devices. Furthermore, we propose a data process-ing algorithm based on the incremental principal component analysis, which can achieve data dimensionality reduction and update of principal components. We also introduce the concept of Squared Prediction Error (SPE) value and realize the abnormal detection of data through the combination of SPE value and data fusion algorithm. To ensure the accuracy and effectiveness of the algorithm, we design a regular-SPE hybrid model update strategy, which enables the principal component to be updated on demand when data anomalies are found. In addition, this strategy can significantly reduce resource consumption growth due to the data analysis architectures. Practical datasets-based simulations have confirmed that the proposed algorithm can perform data fusion and exception processing in real-time on resource-constrained devices; Our model update strategy can reduce the overall system resource consumption while ensuring the accuracy of the algorithm.

• Bulletin of the Korean Chemical Society
• /
• v.29 no.2
• /
• pp.381-388
• /
• 2008

An X-linked skeletal disorder, SEDT (spondyloepiphyseal dysplasia tarda) is a genetic disease characterized by a disproportionately short trunk and short stature caused by mutations in the SEDL gene. This gene is evolutionarily conserved from yeast to human. The yeast SEDL protein ortholog, Trs20p, has been isolated as a member of a large multi-protein complex called the transport protein particle (TRAPP), which is involved in endoplasmic reticulum (ER)-to-Golgi transport. The interaction between SEDL and partner proteins is important in order to understand the molecular mechanism of SEDL functions. We isolated several SEDL-binding proteins derived from rat cells by an immobilized GST-fusion method. Furthermore, the SEDL-homologue proteins were identified using motif based methods. Common motifs between SEDL-binding proteins and SEDL-homologue proteins were classified into seven types and 78 common motifs were revealed. Sequence similarities were contracted to seven types using phylogenetic trees. In general, types I-III and VI were classified as having the function of acetyl-CoA carboxylase, glycogen phosphorylase, isocitrate dehydrogenase, and enolase, respectively, and type IV was found to be functionally related to the GST protein. Types V and VII were found to contribute to TRAPP vesicle trafficking.

• Asian Journal of Atmospheric Environment
• /
• v.7 no.1
• /
• pp.25-37
• /
• 2013

Open biomass burning (excluding biofuels) is an important contributor to air pollution in the Asian region. Estimation of emissions from fires, however, has been problematic, primarily because of uncertainty in the size and location of sources and in their temporal and spatial variability. Hence, more comprehensive tools to estimate wildfire emissions and that can characterize their temporal and spatial variability are needed. Furthermore, an emission processing system that can generate speciated, gridded, and temporally allocated emissions is needed to support air-quality modeling studies over Asia. For these reasons, a biomass-burning emissions modeling system based on satellite imagery was developed to better account for the spatial and temporal distributions of emissions. The BlueSky Framework, which was developed by the USDA Forest Service and US EPA, was used to develop the Asian biomass-burning emissions modeling system. The sub-models used for this study were the Fuel Characteristic Classification System (FCCS), CONSUME, and the Emissions Production Model (EPM). Our domain covers not only Asia but also Siberia and part of central Asia to assess the large boreal fires in the region. The MODIS fire products and vegetation map were used in this study. Using the developed modeling system, biomass-burning emissions were estimated during April and July 2008, and the results were compared with previous studies. Our results show good to fair agreement with those of GFEDv3 for most regions, ranging from 9.7 % in East Asia to 52% in Siberia. The SMOKE modeling system was combined with this system to generate three-dimensional model-ready emissions employing the fire-plume rise algorithm. This study suggests a practicable and maintainable methodology for supporting Asian air-quality modeling studies and to help understand the impact of air-pollutant emissions on Asian air quality.

• International Journal of Advanced Culture Technology
• /
• v.5 no.1
• /
• pp.90-97
• /
• 2017

Nowadays, the most of commercial buildings are build-out with complex architecture and decorated with more complicated interiors of buildings so establishing intelligible escape routes becomes an important case of fire or other emergency in a limited time. The commercial buildings are already equipped with multiple exit signs and these exit signs may create confusion and leads the people into different directions under emergency. This can jeopardize the emergency situation into a chaotic state, especially in a complex layout buildings. There are many research focused on implementing different approached to improve the exit sign system with better visual navigating effects, such as the use of laser beams, the combination of audio and video cues, etc. However the digital signage system based emergency exit sign management is one of the best solution to guide people under emergency situations to escape. This research paper, propose an intelligent evacuation route GUIDE that uses the combination centralized Wireless Sensor Networks (WSN) and digital signage for people safety and avoids dangers from emergency conditions. This proposed system applies WSN to detect the environment condition in the building and uses an evacuation algorithm to estimate the safe route to escape using the sensor information and then activates the signage system to display the safe evacuation route instruction GUIDE according to the location the signage system is installed. This paper presented the prototype of the proposed signage system and execution time to find the route with future research directions. The proposed system provides a natural intelligent evacuation route interface for self or remote operation in facility management to efficiently GUIDE people to the safe exit under emergency conditions.