• Journal of KIISE
• /
• v.43 no.11
• /
• pp.1281-1297
• /
• 2016

In recent years, personal videos have seen a tremendous growth due to the substantial increase in the use of smart devices and networking services in which users create and share video content easily without many restrictions. However, taking both into account would significantly improve event detection performance because videos generally have multiple modalities and the frame data in video varies at different time points. This paper proposes an event detection method. In this method, high-level features are first extracted from multiple modalities in the videos, and the features are rearranged according to time sequence. Then the association of the modalities is learned by means of DNN to produce a personal video event detector. In our proposed method, audio and image data are first synchronized and then extracted. Then, the result is input into GoogLeNet as well as Multi-Layer Perceptron (MLP) to extract high-level features. The results are then re-arranged in time sequence, and every video is processed to extract one feature each for training by means of DNN.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.39 no.10
• /
• pp.1121-1232
• /
• 1990

A hierarchical neural network structure is described for robot PTP trajectory planning. In the first level, the multi-layered Perceptron neural network is used for the inverse kinematics with the back-propagation learning procedure. In the second level, a saccade generation model based joint trajectory planning model in proposed and analyzed with several features. Various simulations are performed to investigate the characteristics of the proposed neural networks.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.8
• /
• pp.2948-2963
• /
• 2015

In the field of computer vision, visual surveillance systems have recently become an important research topic. Growth in this area is being driven by both the increase in the availability of inexpensive computing devices and image sensors as well as the general inefficiency of manual surveillance and monitoring. In particular, the ultimate goal for many visual surveillance systems is to provide automatic activity recognition for events at a given site. A higher level of understanding of these activities requires certain lower-level computer vision tasks to be performed. So in this paper, we propose an intelligent activity recognition model that uses a structure learning method and a classification method. The structure learning method is provided as a K2-learning algorithm that generates Bayesian networks of causal relationships between sensors for a given activity. The statistical characteristics of the sensor values and the topological characteristics of the generated graphs are learned for each activity, and then a neural network is designed to classify the current activity according to the features extracted from the multiple sensor values that have been collected. Finally, the proposed method is implemented and tested by using PETS2013 benchmark data.

• Journal of The Korea Institute of Healthcare Architecture
• /
• v.15 no.2
• /
• pp.51-58
• /
• 2009

These days the trend of special education is changing from special school-based education to special class room based education, and from separated education to integrated education. In accordance with this change, special classes should be planned for multi purposes so that the class room can be used for the place of teaching and learning, guidance, job education. This research surveyed the special teachers working for 937 schools which have special classes(elementary 631, middle 217 high school 89). The result of this survey shows the different responses according to the level of the schools. For education activities, elementary and middle schools put emphasis on curriculum rather than guidance. High education, elementary school should have the places for teaching and learning, student management, play ground. Middle schools give priority to the places for individual learning, computer and practical training. High schools value the places for job education and practical training above for learning.

• Journal of Multimedia Information System
• /
• v.9 no.2
• /
• pp.75-86
• /
• 2022

The existing image quality assessment (IQA) datasets have a small number of samples. Some methods based on transfer learning or data augmentation cannot make good use of image quality-related features. A No Reference (NR)-IQA method based on multi-task training and quality awareness is proposed. First, single or multiple distortion types and levels are imposed on the original image, and different strategies are used to augment different types of distortion datasets. With the idea of weak supervision, we use the Full Reference (FR)-IQA methods to obtain the pseudo-score label of the generated image. Then, we combine the classification information of the distortion type, level, and the information of the image quality score. The ResNet50 network is trained in the pre-train stage on the augmented dataset to obtain more quality-aware pre-training weights. Finally, the fine-tuning stage training is performed on the target IQA dataset using the quality-aware weights to predicate the final prediction score. Various experiments designed on the synthetic distortions and authentic distortions datasets (LIVE, CSIQ, TID2013, LIVEC, KonIQ-10K) prove that the proposed method can utilize the image quality-related features better than the method using only single-task training. The extracted quality-aware features improve the accuracy of the model.

• The Journal of Society for e-Business Studies
• /
• v.26 no.3
• /
• pp.1-20
• /
• 2021

As the Great Reset is discussed at the World Economic Forum due to the COVID-19 pandemic, artificial intelligence, the driving force of the 4th industrial revolution, is also in the spotlight. However, corporate research in the field of artificial intelligence is still scarce. Since 2000, related research has focused on how to create value by applying artificial intelligence to existing companies, and research on how startups seize opportunities and enter among existing businesses to create new value can hardly be found. Therefore, this study analyzed the cases of startups using the comprehensive framework of the multi-level perspective with the research question of how artificial intelligence based startups, a sub-industry of software, have different innovation patterns from the existing software industry. The target firms are gazelle firms that have been certified as venture firms in South Korea, as start-ups within 7 years of age, specializing in machine learning modeling purposively sampled in the medical, finance, marketing/advertising, e-commerce, and manufacturing fields. As a result of the analysis, existing software companies have achieved process innovation from an enterprise-wide integration perspective, in contrast machine learning technology based startups identified unit processes that were difficult to automate or create value by dismantling existing processes, and automate and optimize those processes based on data. The contribution of this study is to analyse the birth of artificial intelligence-based startups and their innovation patterns while validating the framework of an integrated multi-level perspective. In addition, since innovation is driven based on data, the ability to respond to data-related regulations is emphasized even for start-ups, and the government needs to eliminate the uncertainty in related systems to create a predictable and flexible business environment.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.2 no.1
• /
• pp.17-32
• /
• 1992

The multi-layer neural network that has broadly been utilized in designing the controller of robot manipulator possesses the desirable characteristics of learning capacity, by which the uncertain variation of the dynamic parameters of robot can be handled adaptively, and parallel distributed processing that makes it possible to control on real-time. However the error back propagation algorithm that has been utilized popularly in the learning of the multi-layer neural network has the problem of its slow convergence speed. In this paper, an approach to improve the convergence speed is proposed using the fuzzy logic that can effectively handle the uncertain and fuzzy informations by linguistic level. The effectiveness of the proposed algorithm is demonstrated by computer simulation of PUMA 560 robot manupulator.

• Journal of IKEEE
• /
• v.26 no.1
• /
• pp.56-61
• /
• 2022

Various efforts have been made worldwide to respond to environmental problems such as climate change. Research on artificial intelligence (AI)-based energy management has been widely conducted as the most effective way to alleviate the climate change problem. In particular, buildings that account for more than 20% of the total energy delivered worldwide have been focused as a target for energy management using the building energy management system (BEMS). In this paper, we propose a multi-armed bandit (MAB)-based energy management algorithm that can efficiently decide the energy consumption level of the lighting system in each room of the building, while minimizing the discomfort levels of occupants of each room.

• International Journal of Contents
• /
• v.13 no.4
• /
• pp.23-28
• /
• 2017

Of the total economic loss caused by disasters, 40% are due to floods and floods have a severe impact on human health and life. So, it is important to monitor the water level of a river and to issue a flood warning during unfavorable circumstances. In this paper, we propose a modified error function to improve a hydrological modeling using a multi-layer perceptron (MLP) neural network. When MLP's are trained to minimize the conventional mean-squared error function, the prediction performance is poor because MLP's are highly tunned to training data. Our goal is achieved by preventing overspecialization to training data, which is the main reason for performance degradation for rare or test data. Based on the modified error function, an MLP is trained to predict the water level with rainfall data at upper reaches. Through simulations to predict the water level of Nakdong River near a UNESCO World Heritage Site "Hahoe Village," we verified that the prediction performance of MLP with the modified error function is superior to that with the conventional mean-squared error function, especially maximum error of 40.85cm vs. 55.51cm.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.39 no.1
• /
• pp.23-28
• /
• 2021

Data analysis research using deep learning has recently been studied in various field. In this paper, we conduct a GNSS (Global Navigation Satellite System)-based meteorological study applying deep learning by estimating the ZWD (Zenith tropospheric Wet Delay) through MLP (Multi-Layer Perceptron) and LSTM (Long Short-Term Memory) models. Deep learning models were trained with meteorological data and ZWD which is estimated using zenith tropospheric total delay and dry delay. We apply meteorological data not used for learning to the learned model to estimate ZWD with centimeter-level RMSE (Root Mean Square Error) in both models. It is necessary to analyze the GNSS data from coastal areas together and increase time resolution in order to estimate ZWD in various situations.