• 대한임베디드공학회논문지
• /
• 제14권2호
• /
• pp.63-69
• /
• 2019

Recently, ultra-low latency flash storage devices such as Z-SSD and Optane SSD were introduced with the significant technological improvement in the storage devices which provide much faster response time than today's other NVMe SSDs. With such ultra-low latency, $10{\mu}s$, storage devices the cost of context switch could be an overhead during interrupt-driven I/O completion process. As an interrupt-driven I/O completion process could bring an interrupt handling overhead, polling or hybrid-polling for the I/O completion is known to perform better. In this paper, we analyze tail latency problem in a polling process caused by process scheduling in data center environment where multiple applications run simultaneously under one system and we introduce our adaptive polling selection technique which dynamically selects efficient processing method between two techniques according to the system's conditions.

• 대한치과의사협회지
• /
• 제57권5호
• /
• pp.279-287
• /
• 2019

Obstructive sleep apnea syndrome (OSAS) is characterized by obstructive events of the upper airway (UAW) during sleep, which can be associated with clinical signs and symptoms such as snoring, excessive daytime sleepiness, impaired memory, and fatigue. It is associated with many problems like psychosocial problems, physiologic alterations in the cardiovascular and respiratory systems as a result of hypoxia and repeated awakenings during sleep. Conservative treatments such as weight loss, sleep positioning, improvement of sleep hygiene, CPAP and MAD can be performed for the obstructive sleep apnea. However, their effect for the OSA is limited and differs by patient's individual properties. Accordingly, surgical reconstructions of the upper airway must be carried out for the treatment of OSA.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권7호
• /
• pp.2434-2454
• /
• 2021

This paper developed the software testing factory-cloud testing platform (STF-CTP) to address the software compatible issues in various smart devices. Software developers who only require uploading the application under test (AUT) and test script can test plenty of smart devices in STF-CTP. The challenge for the cloud test platform is how to optimize the resource and increase the performance in the limited resource. This paper proposed a new scheduling mechanism and a new process of the system operation which is based on the OpenStack platform. We decrease about 40% memory usage of OpenStack server, increase 3% to 10% Android device usage of STF-CTP, enhance about 80% test job throughput and reduces about 40% test job average waiting time.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2013년도 춘계학술발표대회
• /
• pp.164-165
• /
• 2013

멀티프로세서에서 시스템의 병렬성을 향상시키기 위해서 멀티스레드 프로그램을 이용한다. 이러한 멀티스레드 프로그램은 스레드간 역할을 분담하여 작업을 진행하게 된다. 멀티스레드 프로그램에는 생산자-소비자 구조가 있다. 기존 메모리할당자들은 생산자-소비자 구조에 대한 연구가 진행되지 않고, 크리티컬 섹션이 긴 락을 사용하여 성능상에 문제가 있다. 우리는 이러한 문제점을 독특한 메모리 해제 방법을 통해 해결하였고, 실험을 통해 메모리 할당자의 속도가 향상되는 것을 검증하였다.

• International Journal of Advanced Culture Technology
• /
• 제9권1호
• /
• pp.196-202
• /
• 2021

Currently, as Korea enters into a fast aging society, the problem of dementia population is increasing. In this paper, we intend to contribute to the improvement of welfare for the elderly by developing virtual reality technology and related interface technology to effectively perform hand movements known as effective methods for preventing and treating dementia. As the content of the research and development of this paper, it is designed to be easy for the elderly to use and stimulate brain function by applying VR technology using sensors, and to activate mental and physical activities for the elderly who are marginalized in terms of cultural welfare. We intend to develop by classifying the types of games and contents that can induce them. As a result of this thesis, we developed contents using virtual reality to improve cognitive abilities for elderly people with poor cognitive ability to activate the brains of users' cognition, memory, and attention to prevent and treat dementia I want to contribute.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• pp.3855-3867
• /
• 2022

Name separation (splitting full names into surnames and given names) is not a tedious task in a multiethnic country because the procedure for splitting surnames and given names is ethnicity-specific. Malaysia has multiple main ethnic groups; therefore, separating Malaysian full names into surnames and given names proves a challenge. In this study, we develop a two-phase framework for Malaysian name separation using deep learning. In the initial phase, we predict the ethnicity of full names. We propose a recurrent neural network with long short-term memory network-based model with character embeddings for prediction. Based on the predicted ethnicity, we use a rule-based algorithm for splitting full names into surnames and given names in the second phase. We evaluate the performance of the proposed model against various machine learning models and demonstrate that it outperforms them by an average of 9%. Moreover, transfer learning and fine-tuning of the proposed model with an additional dataset results in an improvement of up to 7% on average.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2010년도 춘계학술발표대회
• /
• pp.68-71
• /
• 2010

임베디드 리눅스의 부팅은 일반적인 경우 Boot Loader, Kernel, Script 로 구성된 초기화 과정, Application Program 의 순서로 이루어진다. 이 경우 부팅 시간은 Power On 에서 최종 Application Program 이 동작을 시작하는 시점까지이다. 따라서 부팅 시간을 줄이는 방법은 부팅 과정의 중간 과정 중 불필요한 과정을 없애거나, 최적화하여 최종 단계에 빠르게 도착하게 만드는 것이다. 이러한 과정들에는 파일시스템 내의 데이터들을 메인 메모리로 복사하는 과정이 포함된다. 임베디드 시스템 내의 파일시스템은 주로 플래시 메모리에 저장되며, 플래시 메모리는 상대적으로 느린 속도로 동작된다. 따라서 부팅 시간은 상당히 많은 부분을 플래시 메모리에서 데이터를 복사하는데 사용된다. 결과적으로 부팅 시간을 줄이는 여러 방법들 중 flash-to-memory copy 의 시간을 줄이는 것은 효율 좋은 방법일 수 있다. 본 논문에서는 임베디드 시스템에 탑재되어 있는 플래시 메모리에서 메모리에 복사시 readahead 를 이용하여 복사시간을 효율화하는 방법을 제안한다.

• International Journal of Advanced Culture Technology
• /
• 제11권1호
• /
• pp.332-336
• /
• 2023

In this paper, we design and implement an ADHD diagnostic rating scale algorithm and virtual reality- based digital treatment by combining virtual reality technology that enables multiple access based on specialized medical data. The VR ADHD digital treatment to be implemented is intended to be applied directly to diagnosis and treatment of child/adolescent ADHD by applying various psychiatric treatment technologies. In order to evaluate the usability of the digital therapeutic agent developed in this paper, an expert group consisting of 2 males and 12 females "targeted counseling psychology majors at the Asian Dementia Center once a day for 3 weeks, ADHD VR-based digital therapeutics for children Experienced and usability evaluation conducted. As a final result of this thesis, it was possible to develop attentional concentration ability, working memory improvement ability, and impulsive control ability in four games of digital therapy, and it was possible to confirm significant results that can be expected to have ADHD therapeutic effects.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2011년도 춘계학술발표대회
• /
• pp.378-381
• /
• 2011

These days rapid improvement in Mobile phones and their multimedia limits made them powerful enough to manage complicated tasks. Image processing related support for mobile devices is extremely comprehensive in wireless telemedicine. A basic challenge is how to get best quality of image with the limited screen size and resources of mobile phones. This paper deals with image processing features (capturing rendering and zooming in and out) of Mobile Media API and Advanced Multimedia Supplements (MMAPI and AMS) developed for Mobile Java Platform and customized algorithm is designed to keep all image task cost efficient by using minimum device resources and memory. This scenario is driven by the need for evaluation of a distant patient that cannot be moved to the expert.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2023년도 춘계학술발표대회
• /
• pp.479-481
• /
• 2023

Generative Adversarial Networks (GANs) have facilitated substantial improvement in single-image super-resolution (SR) by enabling the generation of photo-realistic images. However, the high memory requirements of GAN-based SRs (mainly generators) lead to reduced performance and increased energy consumption, making it difficult to implement them onto resource-constricted devices. In this study, we propose an efficient and compressed architecture for the SR-GAN (generator) model using the model compression technique Knowledge Distillation. Our approach involves the transmission of knowledge from a heavy network to a lightweight one, which reduces the storage requirement of the model by 58% with also an increase in their performance. Experimental results on various benchmarks indicate that our proposed compressed model enhances performance with an increase in PSNR, SSIM, and image quality respectively for x4 super-resolution tasks.