• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.851-853
• /
• 2022

고령화로 인해 65세 이상 운전자가 급증하며 고령운전자의 교통사고 비율이 증가함에 따라 시급한 사회 문제로 떠오르고 있다. 이에 본 연구에서는 객체 검출, 인식 모델을 결합하고 신호등을 인식하여 Text-To-Speech(TTS)로 알리는 쿠버네티스 기반의 프레임워크를 제안한다. 객체 검출 단계에서는 YOLOv5 모델들의 성능을 비교하여 활용하였으며 객체 인식 단계에서는 C-RNN 기반의 attention-OCR 모델을 활용하였다. 이는 신호등의 내부 LED 영역이 아닌 이미지 전체를 인식하는 방식으로 오탐지 요소를 낮춰 인식률을 높였다. 결과적으로 1,628장의 테스트 데이터에서 accuracy 0.997, F1-score 0.991의 성능 평가를 얻어 제안한 프레임워크의 타당성을 입증하였다. 본 연구는 후속 연구에서 특정 도메인에 딥러닝 모델을 한정하지 않고 다양한 분야의 모델을 접목할 수 있도록 하며 고령 운전자 및 신호 위반으로 인한 교통사고 문제를 예방할 수 있다.

• Journal of Bio-Environment Control
• /
• v.32 no.3
• /
• pp.217-225
• /
• 2023

This experiment was conducted to investigate the effects of microwave irradiation on the growth and secondary metabolite contents of lettuce seedlings. Seedlings at three weeks after sowing were treated by a microwave oven for 0, 4, 8, and 12 seconds with 200 W. After cultivation in a close plant production system for 4 weeks, plant growth measurements and secondary metabolite analysis were performed. The results showed that the fresh and dry weights of the shoot and root, leaf area, leaf length, and the number of leaves were decreased as increasing the microwave treatment times. Chlorophyll a and b, total carotenoids were increased and total phenolics were decreased at the 12-second treatment compared to the 4-second treatment. Total flavonoid contents were decreased at the 8-second treatment compared to the control. These results suggest that the changes in the levels of secondary metabolites were caused by oxidative stress. Although there was no significant difference in secondary metabolite contents excluding total flavonoid contents on the microwave treatments compared to the control, the significant difference suggests that the microwave treatment of 200 W and 2.45 GHz may alter secondary metabolite contents of lettuce after 4 weeks.

• Journal of the Korea Convergence Society
• /
• v.11 no.6
• /
• pp.51-57
• /
• 2020

Children with language developmental disabilities often struggle through their lives from a lot of challenges in everyday life and social activities. They're often easily deprived of the opportunity to engage in social activities, because they find difficulty in understanding or using language, a core means of communication. With regard to this issue, AAC(Augmentative and Alternative Communication) can be an effective communication tool for children who are suffering from language disabilities. In this paper, we propose a deep learning-based AI service to make full use of the pictogram as an AAC tool for children with language developmental disabilities to improve not only the ability to interact with others but the capacity to understand language. Using this service, we strive to help these children to more effectively communicate their intention or desire and enhance the quality of life.

• International journal of advanced smart convergence
• /
• v.12 no.4
• /
• pp.260-267
• /
• 2023

Due to the development of science and technology in the 4th Industrial Revolution, a variety of content is being developed and utilized through educational courses linked to digital textbooks. Students use smart devices to engage in realistic virtual learning experiences, interacting with the content in digital textbooks. However, while many realistic contents offer visual and auditory effects like 3D VR, AR, and holograms, olfactory content that evokes actual sensations has not yet been introduced. Therefore, in this paper, we designed and implemented 4D educational content by adding the sense of smell to existing content. This implemented content was tested in classrooms through a curriculum-based evaluation. Classes taught with olfactory-enhanced content showed a higher percentage of correct answers compared to those using traditional audio-visual materials, indicating improved understanding.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.25 no.2
• /
• pp.125-134
• /
• 2022

Recently, there are an increased the number of reports on the misuse or malicious use of an UAS. Thus, many researchers are studying on defense schemes for UAS by developing or improving C-UAS sensor technology. However, the wrong placement of sensors may lead to a defense failure since the proper placement of sensors is critical for UAS defense. In this study, a multi-object optimization model for C-UAS sensor placement in an air base is proposed. To address the issue, we define two objective functions: the intersection ratio of interested area and the minimum detection range and try to find the optimized placement of sensors that maximizes the two functions. C-UAS placement model is designed using a NSGA-II algorithm, and through experiments and analyses the possibility of its optimization is verified.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.4
• /
• pp.186-191
• /
• 2023

Data processing techniques play a critical role in decision-making, including handling missing and outlier data, prediction, and recommendation models. This requires a clear explanation of the validity, reliability, and accuracy of all processes and results. In addition, it is necessary to solve data problems through explainable models using decision trees, inference, etc., and proceed with model lightweight by considering various types of learning. The multi-layer mining classification method that applies the sixth principle is a method that discovers multidimensional relationships between variables and attributes that occur frequently in transactions after data preprocessing. This explains how to discover significant relationships using mining on transactions and model the data through regression analysis. It develops scalable models and logistic regression models and proposes mining techniques to generate class labels through data cleansing, relevance analysis, data transformation, and data augmentation to make explanatory decisions.

• Journal of Animal Science and Technology
• /
• v.63 no.4
• /
• pp.751-758
• /
• 2021

The recessive white (locus c) phenotype observed in chickens is associated with three alleles (recessive white c, albino c^a, and red-eyed white c^re) and causative mutations in the tyrosinase (TYR) gene. The recessive white mutation (c) inhibits the transcription of TYR exon 5 due to a retroviral sequence insertion in intron 4. In this study, we genotyped and sequenced the insertion in TYR intron 4 to identify the mutation causing the unusual white plumage of Yeonsan Ogye chickens, which normally have black plumage. The white chickens had a homozygous recessive white genotype that matched the sequence of the recessive white type, and the inserted sequence exhibited 98% identity with the avian leukosis virus ev-1 sequence. In comparison, brindle and normal chickens had the homozygous color genotype, and their sequences were the same as the wild-type sequence, indicating that this phenotype is derived from other mutation(s). In conclusion, white chickens have a recessive white mutation allele. Since the size of the sample used in this study was limited, further research through securing additional samples to perform validation studies is necessary. Therefore, after validation studies, a selection system for conserving the phenotypic characteristics and genetic diversity of the population could be established if additional studies to elucidate specific phenotype-related genes in Yeonsan Ogye are performed.

• Journal of the Korea Convergence Society
• /
• v.12 no.5
• /
• pp.1-7
• /
• 2021

In the latest trend of machine translation research, the model is pretrained through a large mono lingual corpus and then finetuned with a parallel corpus. Although many studies tend to increase the amount of data used in the pretraining stage, it is hard to say that the amount of data must be increased to improve machine translation performance. In this study, through an experiment based on the mBART model using parallel corpus filtering, we propose that high quality data can yield better machine translation performance, even utilizing smaller amount of data. We propose that it is important to consider the quality of data rather than the amount of data, and it can be used as a guideline for building a training corpus.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.1022-1024
• /
• 2022

최근 인공지능 기술이 급격하게 발전하고 빠르게 보급되면서, 머신러닝 시스템을 대상으로 한 다양한 공격들이 등장하기 시작하였다. 인공지능은 많은 강점이 있지만 인위적인 조작에 취약할 수 있기 때문에, 그만큼 이전에는 존재하지 않았던 새로운 위험을 내포하고 있다고 볼 수 있다. 본 논문에서는 데이터 유형 별 적대적 공격 샘플을 직접 제작하고 이에 대한 효과적인 방어법을 구현하였다. 영상 및 텍스트 데이터를 기반으로 한 적대적 샘플공격을 방어하기 위해 적대적 훈련기법을 적용하였고, 그 결과 공격에 대한 면역능력이 형성된 것을 확인하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.658-661
• /
• 2024

퍼블릭 클라우드 시장이 성장하면서 퍼블릭 클라우드에서 호스팅하는 컴퓨팅 자원으로 구축된 거대하고 복잡한 IT 시스템이 점차 많아지고 있다. 이러한 시스템의 증가는 서비스 장애 발생 확률을 높이므로, 장애 관리 및 선제 감지를 위한 퍼블릭 클라우드 자원의 이상 감지 연구에 대한 수요 또한 증가하고 있다. 그러나 연구에 활용할 수 있는 벤치마크 데이터셋이 없다는 점과, 실제 자원에서 추출할 수 있는 데이터는 레이블링이 되어 있지 않은 불균형 데이터라는 점 때문에 관련 연구가 부족한 상황이다. 이러한 문제를 해결하고자 본 논문은 비지도 방식의 표현 학습 기반 딥러닝 모델을 활용한 이상 감지 시스템을 제안한다. 시스템의 이상 감지 성능을 유지하고자 일정 주기마다 다수의 딥러닝 모델을 재학습하고 비교하여 최적의 모델로 업데이트 하는 방식을 고안하였다. 해당 시스템의 평가에는 실제 퍼블릭 클라우드 자원에서 발생한 메트릭 데이터가 활용됐으며, 그 결과 준수한 이상 감지 성능을 보인다는 것을 확인하였다.