https://scholar.dgist.ac.kr/handle/20.500.11750/189 2026-08-03T09:16:01Z https://scholar.dgist.ac.kr/handle/20.500.11750/60471 Title: Precision enhancement of epidural force-sensing needle with machine learning Author(s): Cho, Gichan; Na, Jongyeol; Lee, Myung Ho; Kwon, Hyun-Jung; Song, Cheol Abstract: Epidural injection is used in pain intervention, requiring precise needle placement within the epidural space. Traditional techniques, such as loss of resistance and fluoroscopy-guided procedures, have limitations, including reliance on subjective assessment and radiation exposure. We proposed an optical force-sensing probe with an offset criterion of the needle tip-distal end to enhance the precision of puncture detection. The offset between the needle tip and the force-sensing probe is adjusted using a piezoelectric motor-based system with feedback position control. A Long Short-Term Memory model is also trained to detect the puncture. Insertion test on silicone phantom and ex-vivo specimens demonstrates that the system's offset range for enhancing precision of puncture detection is between 0.6 mm and 1 mm. Compared to the offset in the previous study, the AUC score of puncture detection increased from 0.61 to 0.86. This approach secures the improvement of puncture detection reliability in robot-assisted epidural injection. 2026-11-30T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60086 Title: 손 떨림 보정 및 광학 거리 제어를 이용한 비접촉 휴대형 공초점 망막 내시현미경 시스템 Author(s): 이명호; 조기찬; 임진택; 나종열; 송철 Abstract: We present a novel handheld confocal endomicroscopy system designed for non-contact, high-resolution retinal imaging, integrating motorized stabilization and tremor prediction. The system combines a custom PZT-driven Lissajous scanning probe, and a common-path swept-source optical coherence tomography (CPSS-OCT) sensor for depth tracking for real-time tremor compensation. This multimodal platform allows continuous imaging at the cellular level without requiring probe-to-tissue contact, addressing critical limitations of conventional probe-based confocal laser endomicroscopy(pCLE) systems. Experimental evaluations using ex-vivo bovine retina samples demonstrated significant enhancements in image clarity, with the CR score increasing by 48.43% under active tremor control. The proposed architecture offers improved image stability, precise axial focusing, and handheld portability, making it a strong candidate for clinical applications in retinal diagnostics and intraoperative imaging. 2025-06-26T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60072 Title: 비전 기반 3차원 자세 추정을 이용한 인간 로봇 충돌 방지 시스템 Author(s): 나종열; 송철 Abstract: As human-robot collaboration becomes increasingly prevalent in medical and industrial settings, the risk of collisions is rising due to shared workspaces. Traditional approaches rely on stopping mechanisms or external sensors, which may limit system adaptability. This study proposes a camera-based method utilizing a convolutional neural network (CNN) to estimate the 3D joint coordinates of robots and humans within a unified camera coordinate frame. It estimates the 3D positions of the robot's seven joints with an average error below 5.5 mm, enabling real-time collision risk assessment through joint distance computation. This vision-based approach provides precise spatial awareness without requiring additional external sensors, enhancing safety and operational efficiency. 2025-06-25T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/59994 Title: Handheld Confocal Endomicroscope System with Tremor Compensation for Retinal Imaging Author(s): Lee, Myung Ho; Cho, Gichan; Im, Jintaek; Na, Jongyeol; Song, Cheol Abstract: Advancements in biophotonics have driven the development of miniaturized imaging probes for high-resolution in vivo imaging. Probe-based confocal laser endomicroscopy (pCLE) enables cellular-level visualization of tissues but remains challenging for retinal imaging due to the need for non-contact operation, tremor compensation, and precise focal control. This study introduces a novel handheld confocal endomi-croscope system that integrates a custom-built imaging probe, an optical coherence tomography (OCT) distance sensor, and motor-assisted tremor suppression to improve imaging stability and resolution. The system employs a fiber-based common-path swept-source OCT (CPSS-OCT) sensor to maintain a stable focal distance while compensating for involuntary hand tremors using motorized stabilization. A gated recurrent unit (GRU)-based tremor prediction algorithm further enhances image stability. The imaging probe features a PZT tube-driven fiber cantilever resonance for Lissajous scanning, providing a wide field of view with minimal image distortion. In experiments using bovine eye samples, the CR score improved from 0.318 to 0.472, with a 48.43% increase in the in-focus condition when tremor compensation was activated, confirming enhanced image clarity and stability. Experimental results demonstrate that the system effectively stabilizes imaging, reduces motion artifacts, and ensures high-resolution, non-contact retinal imaging. By addressing the limitations of conventional pCLE devices, this system represents a significant advancement in ophthalmic imaging and can potentially improve retinal diagnostics and precision-guided interventions. 2025-10-21T15:00:00Z