http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : Guebin Hwang(황규빈) (검색결과 2 건)
- 무료
- 기관 내 무료
- 유료
-
Nostril Detection Algorithms for Visual Control of Automatic COVID-19 Swab Sampling Robot Systems
Guebin Hwang(황규빈),Sungwook Yang(양성욱) 대한기계학회 2021 대한기계학회 춘추학술대회 Vol.2021 No.4
Coronavirus disease 2019 (COVID-19) has spread rapidly and become a global pandemic. Initial screening of patients is most critical to prevent the spread of COVID-19. However, nasopharyngeal swab regarded as a reference sampling method to detect COVID-19 may pose a high risk of cross-infection and also the high workload of medical professionals. We believe that fully automatic sample collection robot systems would have great potential to address these issues. Therefore, we propose a robot system capable automatically of inserting a sampling swab through one nostril using a camera attached to the system. This study aims to evaluate the performance of nostril detection algorithms based on deep learning, Faster RCNN with a two-stage detector and YOLOv3 with a one-stage detector. For a testing set of 100-nostril photos, Faster RCNN showed a detection accuracy of 98% within 0.30-s processing time, whereas YOLOv3 was with 100%-accuracy within 0.14 s. Both deep-learning based approaches outperform a conventional machine learning algorithm, Viola-Jones algorithm, effective for face detection. Although the Viola-Jones algorithm reduced the processing time to 0.03 s, only an accuracy of 24% was obtained for the data set. It concludes that YOLOv3 is suitable for the real-time visual servo control of the COVID-19 sampling robot in terms of accuracy and speed. Future work includes real-time nostril tracking and visual servo control to accomplish fully automated swab sampling.
-
Visual Control of Automated COVID-19 Swab Sampling Robot
Guebin Hwang(황규빈),Jongwoon Lee(이종원),Sungwook Yang(양성욱) 대한기계학회 2021 대한기계학회 춘추학술대회 Vol.2021 No.11
As coronavirus disease 2019 (COVID-19) has become a global pandemic, the initial screening of patients is regarded as the most effective way to prevent the spread of COVID-19. However, enormous swap sampling has led to the high workload of medical professionals as the number of infection has increased. Therefore, we propose a swab sampling robot systems in a fully automated fashion to address these issues. For automation of nasopharyngeal swab, the proposed robot system is designed to automatically insert a sampling swab through one nostril and collect sample via visual feedback. A prototype of the robot system incorporates a 6-DOF manipulator and a RGB-D sensor to investigate the feasibility of a visual-servo control scheme in automatic swab sampling. For accurate and safe operation, a deep learning-based nostril detection algorithm is adopted, which offers an accuracy of 99% at 60 FPS. Given the nostril position in 3D, the robot is planned to reach the target within a specific time. Since the kinematic control of the robot is prone to failure in reaching the target, we introduce the visual-servo control of the robot via the detection and tracking of error between the target position and the current swab tip. As a result, the visual-servo controlled robot could successfully reach the nostril target within an error of 1.11 mm on average for 30 trails trials different swabs, while the non-visual control with a 10.58-mm positioning error failed to reach the target.
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
