• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.12
• /
• pp.411-418
• /
• 2020

In the field of information security, IDS(Intrusion Detection System) is normally classified in two different categories: signature-based IDS and anomaly-based IDS. Many studies in anomaly-based IDS have been conducted that analyze network traffic data generated in cyberspace by machine learning algorithms. In this paper, we studied pre-processing methods to overcome performance degradation problems cashed by rare classes. We experimented classification performance of a Machine Learning algorithm by reconstructing data set based on rare classes and semi rare classes. After reconstructing data into three different sets, wrapper and filter feature selection methods are applied continuously. Each data set is regularized by a quantile scaler. Depp neural network model is used for learning and validation. The evaluation results are compared by true positive values and false negative values. We acquired improved classification performances on all of three data sets.

• Journal of Advanced Technology Convergence
• /
• v.1 no.2
• /
• pp.01-09
• /
• 2022

Recently, safety issues in companies and public institutions are no longer a task that can be postponed, and when a major safety accident occurs, not only direct financial loss, but also indirect loss of social trust in the company and public institution is greatly increased. In particular, in the case of a fatal accident, the damage is even more serious. Accordingly, as companies and public institutions expand their investments in industrial safety education and prevention, open AI learning model creation technology that enables safety management services without being affected by user behavior in industrial sites where high-risk situations exist, edge terminals System development using inter-AI collaboration technology, cloud-edge terminal linkage technology, multi-modal risk situation determination technology, and AI model learning support technology is underway. In particular, with the development and spread of artificial intelligence technology, research to apply the technology to safety issues is becoming active. Therefore, in this paper, an open cloud platform design method that can support AI model learning for high-risk site safety management is presented.

• Journal of Biomedical Engineering Research
• /
• v.42 no.5
• /
• pp.241-249
• /
• 2021

Recently, artificial intelligence for diagnosis system of obstetric diseases have been actively studied. Artificial intelligence diagnostic assist systems, which support medical diagnosis benefits of efficiency and accuracy, may experience problems of poor learning accuracy and reliability when inappropriate images are the model's input data. For this reason, before learning, We proposed an algorithm to exclude unread cervical imaging. 2,000 images of read cervical imaging and 257 images of unread cervical imaging were used for this study. Experiments were conducted based on the statistical method Radiomics to extract feature values of the entire images for classification of unread images from the entire images and to obtain a range of read threshold values. The degree to which brightness, blur, and cervical regions were photographed adequately in the image was determined as classification indicators. We compared the classification performance by learning read cervical imaging classified by the algorithm proposed in this paper and unread cervical imaging for deep learning classification model. We evaluate the classification accuracy for unread Cervical imaging of the algorithm by comparing the performance. Images for the algorithm showed higher accuracy of 91.6% on average. It is expected that the algorithm proposed in this paper will improve reliability by effectively excluding unread cervical imaging and ultimately reducing errors in artificial intelligence diagnosis.

• Korean Journal of Cognitive Science
• /
• v.26 no.4
• /
• pp.377-392
• /
• 2015

Korean semantic role labeling (SRL) is usually performed by a machine learning and requires a lot of corpus. However, the Korean PropBank used in Korean SRL system is less than PropBank. It leads to a low performance. Therefore, we expand the annotated corpus and verb frames for Korean SRL system to expand the Korean PropBank corpus. Most of the SRL system have a domain-dependent performance so, the performance may decrease if domain was changed. In this paper, we use the domain adaptation technique to reduce decreasing performance with the existing corpus and the small size of new domain corpus. We apply the domain adaptation technique to Structural SVM and Deep Neural Network. The experimental result show the effectiveness of the domain adaptation technique.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.4
• /
• pp.34-40
• /
• 2018

This paper presents a new method to estimate the pose of a moving object accurately using a monocular camera and a low-end GPS+IMU sensor system. For this goal, we adopted a deep neural network for the semantic segmentation of input images and compared the results with a semantic map of a neighborhood. In this map matching, we use weight tables to deal with label inconsistency effectively. Signals from a low-end GPS+IMU sensor system are used to limit search spaces and minimize the proposed function. For the evaluation, we added noise to the signals from a high-end GPS-IMU system. The results show that the pose can be recovered from the noisy signals. We also show that the proposed method is effective in handling non-open-sky situations.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.1
• /
• pp.169-174
• /
• 2018

There are many CCTV cameras connected to the video surveillance and monitoring center for the safety of citizens, and it is difficult for a few monitoring agents to monitor many channels of videos. In this paper, we propose an intelligent video surveillance system utilizing a safety map to efficiently monitor many channels of CCTV camera videos. The safety map establishes the frequency of crime occurrence as a database, expresses the degree of crime risk and makes it possible for agents of the video surveillance center to pay attention when a woman enters the crime risk area. The proposed gender classification method is processed in the order of pedestrian detection, tracking and classification with deep training. The pedestrian detection and tracking uses Adaboost algorithm and probabilistic data association filter, respectively. In order to classify the gender of the pedestrian, relatively simple AlexNet is applied to determine gender. Experimental results show that the proposed gender classification method is more effective than the conventional algorithm. In addition, the results of implementation of intelligent video security system combined with safety map are introduced.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.6
• /
• pp.1-8
• /
• 2021

In this paper, we propose reinforced VGG style network structure for low performance embedded system to classify low resolution infrared image. The combination of reinforced VGG style network structure and global average pooling makes lower computational complexity and higher accuracy. The proposed method classify the synthesize image which have 9 class 3,723,328ea images made from OKTAL-SE tool. The reinforced VGG style network structure composed of 4 filters on input and 16 filters on output from max pooling layer shows about 34% lower computational complexity and about 2.4% higher accuracy then the first parameter minimized network structure made for embedded system composed of 8 filters on input and 8 filters on output from max pooling layer. Finally we get 96.1% accuracy model. Additionally we confirmed the about 31% lower inference lead time in ported C code.

• Information Systems Review
• /
• v.23 no.4
• /
• pp.69-86
• /
• 2021

As artificial intelligence technology advances, personalized recommendation systems using big data have attracted huge attention. In the case of beauty products, product preferences are clearly divided depending on customers' skin types and sensitivity along with individual tastes, so it is necessary to provide customized recommendation services based on accumulated customer data. Therefore, by employing deep learning methods, this study proposes a neural network-based recommendation model utilizing both product search history and context information such as gender, skin types and skin worries of customers. The results show that our model with context information outperforms collaborative filtering-based recommender system models using customer search history.

• Annual Conference of KIPS
• /
• 2017.11a
• /
• pp.845-848
• /
• 2017

In this paper, we implemented a system to detect dangerous situations by recognizing the dangerous elements for infants by reading 2D images of children's houses, parks, playgrounds, and living rooms where infants are present through Faster R-CNN. We have implemented a detection model based on data that can be easily obtained from real life. Currently, machine learning is commercialized based on speech recognition and behavior data. However, this model can be applied to various service fields Respectively.

• The Journal of Korea Robotics Society
• /
• v.16 no.3
• /
• pp.276-282
• /
• 2021

The system for recognizing text in natural scenes has been applied in various industries. However, due to the change in brightness that occurs in nature such as light reflection and shadow, the text recognition performance significantly decreases. To solve this problem, we propose an adversarial shadow generation and training algorithm that is robust to shadow changes. The adversarial shadow generation and training algorithm divides the entire image into a total of 9 grids, and adjusts the brightness with 4 trainable parameters for each grid. Finally, training is conducted in a adversarial relationship between the text recognition model and the shaded image generator. As the training progresses, more and more difficult shaded grid combinations occur. When training with this curriculum-learning attitude, we not only showed a performance improvement of more than 3% in the ICDAR2015 public benchmark dataset, but also confirmed that the performance improved when applied to our's android application text recognition dataset.