http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
콜라주 기법에 의한 비디오 생성을 위한 탐색적 실험 분석
조형래(Hyeongrae Cho),박구만(Gooman Park) 한국방송·미디어공학회 2020 한국방송공학회 학술발표대회 논문집 Vol.2020 No.11
딥러닝이 정답을 찾아가는 연구과정이라면 미술은 정답이나 오답의 단정적 결과보다는 미추(아름다움과 추함)를 포함하는 과정적, 창조적 행위에 가깝다고 할 수 있다. 다시 말하면 미술은 0과 1로만 환원할 수 없는 세계를 기술하여 감동을 주는 유기적 규칙이 내재되어 있고 때로는 과학이 만들어낸 결론을 뒤집는 반상식적 추론을 하기도 한다. 그러므로 딥러닝은 예술적 방식을 통하여 과학의 상식적 추론과의 좋은 거리(Fine distance)를 유지할 필요성이 있는데, 이를 위해서 기존 딥러닝의 이미지 생성과 관련하여 Distance, Classification, Optimization 등의 문제를 미술 표현 기법과 목적이 담겨있는 창작자의 Statement 키워드와의 유사성과 차이점을 비교 분석할 필요가 있다고 생각한다. 시각적 표현과 관련된 딥러닝의 성능은 아직 사람의 표현능력에 못 미치고 있어 본 논문에서는 콜라주 기법에 의한 비디오 생성을 위한 탐색적 실험 분석을 목적으로 GAN을 활용한 콜라주 비디오를 제작하고 그 문제점과 개선점을 제안하고자 한다.
조형래(Hyeongrae Cho),박구만(Gooman Park) 한국방송·미디어공학회 2021 방송공학회논문지 Vol.26 No.1
In the field of deep learning, there are many algorithms mainly after GAN in research related to generation, but in terms of generation, there are similarities and differences with art. If the generation in the engineering aspect is mainly to judge the presence or absence of a quantitative indicator or the correct answer and the incorrect answer, the creation in the artistic aspect creates a creation that interprets the world and human life by cross-validating and doubting the correct answer and incorrect answer from various perspectives. In this paper, the video generation ability of deep learning was interpreted from the perspective of collage and compared with the results made by the artist. The characteristic of the experiment is to compare and analyze how much GAN reproduces the result of the creator made with the collage technique and the difference between the creative part, and investigate the satisfaction level by making performance evaluation items for the reproducibility of GAN. In order to experiment on how much the creators statement and purpose of expression were reproduced, a deep learning algorithm corresponding to the statement keyword was found and its similarity was compared. As a result of the experiment, GAN did not meet much expectations to express the collage technique. Nevertheless, the image association showed higher satisfaction than human ability, which is a positive discovery that GAN can show comparable ability to humans in terms of abstract creation.
스마트워치의 생리 데이터 기반 질병 모니터링 체계를 활용한 COVID-19 이상탐지
김진현(Jin Hyun Kim),한용섭(Yong Seop Han),조형래(Hyeongrae Cho),윤혜린(Hyerin Yoon),김현수(Hyeonsu Kim),구다예(Daye Gu),강태신(Taeseen Kang) 대한전기학회 2021 전기학회논문지 Vol.70 No.8
Real-Time vital-sign from patients are important information that implies the current health status and behavior of patients. Recently, Mishra et al. have shown that COVID-19 can be detected by analyzing the patient’s vital signs and behaviors, i.e., heart rates and steps, using anomaly detection techniques. This paper presents a medical IoT platform, called MiT Eco-platform, which is designed to gather patient’s physiological data through a smartwatch and to increase the efficiency of data labeling for building an AI model for medical diagnosis and treatment. Furthermore, we present a real-time COVID-19 detection approach advanced from the approach of using anomaly detection Mishra et al. that will be run on MiT Eco-platform. As a result, we show performance evaluation results of preemptively detecting the COVID-19 infection for the same samples of the COVID-19 infected ones of Mishra et al., comparing with the anomaly detection approach of Mishra et al.. We expect that physiological data through smartwatches on daily life can be continuously gathered and effectively labeled by the MiT Eco-platform for various studies in medical area.