• Journal of IKEEE
• /
• v.24 no.1
• /
• pp.147-153
• /
• 2020

Conventional lane detection algorithms have problems in that the detection rate is lowered in road environments having a large change in curvature and illumination. The probabilistic Hough transform method has low lane detection rate since it exploits edges and restrictive angles. On the other hand, the method using a sliding window can detect a curved lane as the lane is detected by dividing the image into windows. However, the detection rate of this method is affected by road slopes because it uses affine transformation. In order to detect lanes robustly and avoid obstacles, we propose driving assist system using semantic segmentation based on deep learning. The architecture for segmentation is SegNet based on VGG-16. The semantic image segmentation feature can be used to calculate safety space and predict collisions so that we control a vehicle using adaptive-MPC to avoid objects and keep lanes. Simulation results with CARLA show that the proposed algorithm detects lanes robustly and avoids unknown obstacles in front of vehicle.

• Journal of the Korea Institute of Building Construction
• /
• v.21 no.5
• /
• pp.397-408
• /
• 2021

The construction industry has the highest occupational accidents/injuries and has experienced the most fatalities among entire industries. Korean government installed surveillance camera systems at construction sites to reduce occupational accident rates. Construction safety managers are monitoring potential hazards at the sites through surveillance system; however, the human capability of monitoring surveillance system with their own eyes has critical issues. A long-time monitoring surveillance system causes high physical fatigue and has limitations in grasping all accidents in real-time. Therefore, this study aims to build a deep learning-based safety monitoring system that can obtain information on the recognition, location, identification of workers and heavy equipment in the construction sites by applying multiple object tracking with instance segmentation. To evaluate the system's performance, we utilized the Microsoft common objects in context and the multiple object tracking challenge metrics. These results prove that it is optimal for efficiently automating monitoring surveillance system task at construction sites.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.12
• /
• pp.123-132
• /
• 2021

Malicious activities of Botnets are responsible for huge financial losses to Internet Service Providers, companies, governments and even home users. In this paper, we try to confirm the possibility of detecting botnet traffic by applying the deep learning model Convolutional Neural Network (CNN) using the CTU-13 botnet traffic dataset. In particular, we classify three classes, such as the C&C traffic between bots and C&C servers to detect C&C servers, traffic generated by bots other than C&C communication to detect bots, and normal traffic. Performance metrics were presented by accuracy, precision, recall, and F1 score on classifying both known and unknown botnet traffic. Moreover, we propose a stackable botnet detection system that can load modules for each botnet type considering scalability and operability on the real field.

• Journal of the Korea Computer Graphics Society
• /
• v.25 no.3
• /
• pp.1-9
• /
• 2019

The image obtained from systems such as autonomous driving cars or fire-fighting robots often suffer from several degradation such as noise, motion blur, and compression artifact due to multiple factor. It is difficult to apply image recognition to these degraded images, then the image restoration is essential. However, these systems cannot recognize what kind of degradation and thus there are difficulty restoring the images. In this paper, we propose the deep neural network, which restore natural images from images degraded in several ways such as noise, blur and JPEG compression in situations where the distortion applied to images is not recognized. We adopt the channel attention modules and skip connections in the proposed method, which makes the network focus on valuable information to image restoration. The proposed method is simpler to train than other methods, and experimental results show that the proposed method outperforms existing state-of-the-art methods.

• Korean Journal of Remote Sensing
• /
• v.35 no.2
• /
• pp.279-288
• /
• 2019

We evaluated the land cover classification performance of SegNet, which features semantic segmentation of aerial imagery. We selected four semantic classes, i.e., urban, farmland, forest, and water areas, and created 2,000 datasets using aerial images and land cover maps. The datasets were divided at a 8:2 ratio into training (1,600) and validation datasets (400); we evaluated validation accuracy after tuning the hyperparameters. SegNet performance was optimal at a batch size of five with 100,000 iterations. When 200 test datasets were subjected to semantic segmentation using the trained SegNet model, the accuracies were farmland 87.89%, forest 87.18%, water 83.66%, and urban regions 82.67%; the overall accuracy was 85.48%. Thus, deep learning-based semantic segmentation can be used to classify land cover.

• Database Research
• /
• v.34 no.3
• /
• pp.3-13
• /
• 2018

Text classification is one of the text mining technologies that classifies a given textual document into its appropriate categories and is used in various fields such as spam email detection, news classification, question answering, emotional analysis, and chat bot. In general, the text classification system utilizes machine learning algorithms, and among a number of algorithms, naïve Bayes and support vector machine, which are suitable for text data, are known to have reasonable performance. Recently, with the development of deep learning technology, several researches on applying deep neural networks such as recurrent neural networks (RNN) and convolutional neural networks (CNN) have been introduced to improve the performance of text classification system. However, the current text classification techniques have not yet reached the perfect level of text classification. This paper focuses on the fact that the text data is expressed as a vector only with the word dimensions, which impairs the semantic information inherent in the text, and proposes a neural network architecture based upon the semantic tensor space model.

• Journal of Platform Technology
• /
• v.9 no.3
• /
• pp.44-51
• /
• 2021

In general, golden time refers to the most important time in the initial response to accidents such as saving lives or extinguishing fires. The golden time varies from disaster to disaster, but is aimed at five minutes in terms of fire and first aid. However, for the actual site, the average dispatch time for ambulances is 9 minutes and the average transfer time is 17.6 minutes, which is quite large compared to the golden time. There are various causes for this delay, but the main cause is traffic jams. In order to solve the problem, the government has established emergency car concession obligations and secured golden time to prioritize ambulances in places with the highest accident rate, but it is not a solution in rush hour when traffic is increasing rapidly. Therefore, this paper proposed a deep learning-based emergency vehicle priority signal system using collected sound data by installing sound sensors on traffic lights and conducted an experiment to classify frequency signals that differ depending on the distance of the emergency vehicle.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.6
• /
• pp.115-121
• /
• 2020

With the spread of IoT technology, various IoT applications using facial recognition are emerging. This paper describes the design and implementation of a remote control system using deep learning-based face recognition and hand gesture recognition. In general, an application system using face recognition consists of a part that takes an image in real time from a camera, a part that recognizes a face from the image, and a part that utilizes the recognized result. Raspberry PI, a single board computer that can be mounted anywhere, has been used to shoot images in real time, and face recognition software has been developed using tensorflow's FaceNet model for server computers and hand gesture recognition software using OpenCV. We classified users into three groups: Known users, Danger users, and Unknown users, and designed and implemented an application that opens automatic door locks only for Known users who have passed both face recognition and hand gestures.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.4
• /
• pp.71-82
• /
• 2021

In this study, a novel method for monitoring road pavements using the Mobile Mapping System (MMS) and a deep learning crack detection system was presented. Furthermore, an optimal maintenance method through economic analysis was presented targeting the pavement section of Sejong City. As a result of monitoring the pavement conditions, it was confirmed that the pavement ratings were good in the order of national highways, municipal roads, and roads of provinces. In addition, economic analysis using the pavement deterioration model showed that micro-surfacing, one of the preventive maintenance methods, is the most economical in terms of maintenance costs and user benefits. The results of this study are expected to be used as fundamental reference for local governments' infrastructure management plans.

• Journal of Navigation and Port Research
• /
• v.46 no.4
• /
• pp.344-350
• /
• 2022

Due to the increase in container cargo volume, the congestion of container terminals is increasing and the waiting time of gate in-out trucks has significantly lengthened at container yards and gates, resulting in severe inefficiency in gate in-out truck operations as well as port operations. To resolve this problem, the Busan Port Authority and terminal operator provide services such VBS, terminal congestion information, and expected operation processing time information. However, the visible effect remains insufficient, as it may differ from actual waiting time.. Thus, as basic data to resolve this problem, this study presents deep learning based average gate in-out truck waiting time prediction models, using container gate in-out information at Busan New Port. As a result of verifying the predictive rate through comparison with the actual average waiting time, it was confirmed that the proposed predictive models showed high predictive rate.