• International journal of advanced smart convergence
• /
• v.9 no.1
• /
• pp.113-120
• /
• 2020

Recently, the license plate OCR system has been commercialized in a variety of fields and preferred utilizing low-cost embedded systems using only cameras. This system has a high recognition rate of about 98% or more for the environments such as parking lots where non-vehicle is restricted; however, the environments where non-vehicle objects are not restricted, the recognition rate is about 50% to 70%. This low performance is due to the changes in the environment by non-vehicle objects in real-time situations that occur anomaly data which is similar to the license plates. In this paper, we implement the appropriate anomaly detection based on semi-supervised learning for the license plate OCR system in the real-time environment where the appearance of non-vehicle objects is not restricted. In the experiment, we compare systems which anomaly detection is not implemented in the preceding research with the proposed system in this paper. As a result, the systems which anomaly detection is not implemented had a recognition rate of 77%; however, the systems with the semi-supervised learning based on anomaly detection had 88% of recognition rate. Using the techniques of anomaly detection based on the semi-supervised learning was effective in detecting anomaly data and it was helpful to improve the recognition rate of real-time situations.

• Electronics and Telecommunications Trends
• /
• v.36 no.4
• /
• pp.145-153
• /
• 2021

Data imputation is a crucial issue in data analysis because quality data are highly correlated with the performance of AI models. Particularly, it is difficult to collect quality time-series data for uncertain situations (for example, electricity blackout, delays for network conditions). Thus, it is necessary to research effective methods of time-series data imputation. Many studies on time-series data imputation can be divided into 5 parts, including statistical based, matrix-based, regression-based, deep learning (RNN and GAN) based methodologies. This study reviews and organizes these methodologies. Recently, deep learning-based imputation methods are developed and show excellent performance. However, it is associated to some computational problems that make it difficult to use in real-time system. Thus, the direction of future work is to develop low computational but high-performance imputation methods for application in the real field.

• Journal of information and communication convergence engineering
• /
• v.18 no.1
• /
• pp.28-32
• /
• 2020

Railway transportation is the main land-based transportation in most countries. Accordingly, railway-transportation safety has always been a key issue for many researchers. Railway pedestrian accidents are the main reasons of railway-transportation casualties. In this study, we conduct experiments to determine which of the latest convolutional neural network models and algorithms are appropriate to build pedestrian railroad accident prevention systems. When a drone cruises over a pre-specified path and altitude, the real-time status around the rail is recorded, following which the image information is transmitted back to the server in time. Subsequently, the images are analyzed to determine whether pedestrians are present around the railroads, and a speed-deceleration order is immediately sent to the train driver, resulting in a reduction of the instances of pedestrian railroad accidents. This is the first part of an envisioned drone-based intelligent security system. This system can effectively address the problem of insufficient manual police force.

• Journal of Korea Multimedia Society
• /
• v.24 no.11
• /
• pp.1481-1491
• /
• 2021

Since most biosignals rely on contact-based measurement, there is still a problem in that it is hard to provide convenience to users by applying them to daily life. In this paper, we present a mobile application for estimating heart rate based on a deep learning model. The proposed application measures heart rate by capturing real-time face images in a non-contact manner. We trained a three-dimensional convolutional neural network to predict photoplethysmography (PPG) from face images. The face images used for training were taken in various movements and situations. To evaluate the performance of the proposed system, we used a pulse oximeter to measure a ground truth PPG. As a result, the deviation of the calculated root means square error between the heart rate from remote PPG measured by the proposed system and the heart rate from the ground truth was about 1.14, showing no significant difference. Our findings suggest that heart rate measurement by mobile applications is accurate enough to help manage health during daily life.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.483-486
• /
• 2020

Age estimation and gender classification for human is a classic problem in computer vision. Almost research focus just only one task and the models are too heavy to run on low-cost system. In our research, we aim to apply multitasking learning to perform both task on a lightweight model which can achieve good precision on embedded system in the real time.

• Journal of Information Technology Services
• /
• v.20 no.3
• /
• pp.75-86
• /
• 2021

As artificial intelligence is actively studied, chatbot systems are being applied to various fields. In particular, many chatbot systems for psychological counseling have been studied that can comfort modern people. However, while most psychological counseling chatbots are studied as rule-base and deep learning-based chatbots, there are large limitations for each chatbot. To overcome the limitations of psychological counseling using such chatbots, we proposes a novel psychological counseling AI chatbot system. The proposed system consists of a GPT-2 model that generates output sentence for Korean input sentences and an Electra model that serves as sentiment analysis and anxiety cause classification, which can be provided with psychological tests and collective intelligence functions. At the same time as deep learning-based chatbots and conversations take place, sentiment analysis of input sentences simultaneously recognizes user's emotions and presents psychological tests and collective intelligence solutions to solve the limitations of psychological counseling that can only be done with chatbots. Since the role of sentiment analysis and anxiety cause classification, which are the links of each function, is important for the progression of the proposed system, we experiment the performance of those parts. We verify the novelty and accuracy of the proposed system. It also shows that the AI chatbot system can perform counseling excellently.

• The Journal of Korea Robotics Society
• /
• v.18 no.2
• /
• pp.225-232
• /
• 2023

With the recent increase of multiple robots cooperating in smart manufacturing logistics environments, it has become very important how to predict the safety and efficiency of the individual tasks and dynamically assign them to the best one of available robots. In this paper, we propose a novel task policy learner based on deep relational reinforcement learning for predicting the safety and efficiency of tasks in a multi-robot manufacturing logistics environment. To reduce learning complexity, the proposed system divides the entire safety/efficiency prediction process into two distinct steps: the policy parameter estimation and the rule-based policy inference. It also makes full use of domain-specific knowledge for policy rule learning. Through experiments conducted with virtual dynamic manufacturing logistics environments using NVIDIA's Isaac simulator, we show the effectiveness and superiority of the proposed system.

• Journal of Advanced Navigation Technology
• /
• v.24 no.6
• /
• pp.620-626
• /
• 2020

This research propose and simulate a solar power generation system monitoring system based on Modbus TCP communication using RaspberryPi, an IOT equipment, as a master and an inverter as a slave. In this model, various sensors are added to the RaspberryPi to add necessary information for monitoring solar power plants, and power generation prediction and monitoring information are transmitted to the smart phone through real-time power generation prediction. In addition, information that is continuously generated by the solar power plant is built on the server as big data, and a deep learning model for predicting power generation is trained and updated. As a result of the study, stable communication was possible based on Modbus TCP with the Raspberry Pi in the inverter, and real-time prediction was possible with the deep learning model learned in the Raspberry Pi. The server was able to train various deep learning models with big data, and it was confirmed that LSTM showed the best error with a learning error of 0.0069, a test error of 0.0075, and an RMSE of 0.0866. This model suggested that it is possible to implement a real-time monitoring system that is simpler, more convenient, and can predict the amount of power generation for inverters of various manufacturers.

• Smart Media Journal
• /
• v.11 no.1
• /
• pp.31-37
• /
• 2022

This paper proposes a study on applying signal processing and deep learning for sound recognition that detects sounds commonly heard in daily life (Screaming, Clapping, Crowd_clapping, Car_passing_by and Back_ground, etc.). In the proposed sound recognition, several techniques related to the spectrum of sound waves, augmentation of sound data, ensemble learning for various predictions, convolutional neural networks (CNN) deep learning, and two-dimensional (2-D) data are used for improving the recognition accuracy. The proposed sound recognition technology shows that it can accurately recognize various sounds through experiments.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.10
• /
• pp.373-380
• /
• 2022

This paper presents a deep learning based group synchronization that supports networked immersive interactions between remote users. The goal of group synchronization is to enable all participants to synchronously interact with others for increasing user presence Most previous methods focus on NTP-based clock synchronization to enhance time accuracy. Moving average filters are used to control media playout time on the synchronization server. As an example, the exponentially weighted moving average(EWMA) would be able to track and estimate accurate playout time if the changes in input data are not significant. However it needs more time to be stable for any given change over time due to codec and system loads or fluctuations in network status. To tackle this problem, this work proposes the Deep Group Synchronization(DeepGroupSync), a group synchronization based on deep learning that models important features from the data. This model consists of two Gated Recurrent Unit(GRU) layers and one fully-connected layer, which predicts an optimal playout time by utilizing the sequential playout delays. The experiments are conducted with an existing method that uses the EWMA and the proposed method that uses the DeepGroupSync. The results show that the proposed method are more robust against unpredictable or rapid network condition changes than the existing method.