https://scholar.dgist.ac.kr/handle/20.500.11750/91 2026-07-11T19:00:55Z https://scholar.dgist.ac.kr/handle/20.500.11750/60432 Title: 문맥 인지 비디오 인스턴스 세그먼테이션 방법 Author(s): 이승훈; 최민우; 서지완; 한길준; 임성훈 https://scholar.dgist.ac.kr/handle/20.500.11750/60427 Title: 단안 카메라 이미지에 대한 깊이 추정 방법 Author(s): 최원혁; 임성훈; 신민규 Abstract: 본 개시의 일 실시예의 컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램이 개시된다. 상기 컴퓨터 프로그램은, 하나 이상의 프로세서에서 실행되는 경우 객체의 깊이 추정을 위한 이하의 방법들을 수행하도록 하며, 상기 방법은, 피처 추출 네트워크에서 학습용 이미지 데이터에 포함된 둘 이상의 객체들 각각을 피처 공간에 피처 점(feature point)으로 매핑하는 단계; 상기 피처 공간에 매핑 된 각각의 피처 점들 중 적어도 일부의 피처 점들 사이의 피처 공간에서의 거리와 상기 둘 이상의 객체들 중 상기 피처 점들과 대응되는 적어도 일부의 객체들 사이의 깊이 공간에서의 거리를 비교하는 단계; 및 상기 비교 결과에 기초하여, 상기 피처 점들 사이의 피처 공간에서의 거리와 상기 피처 점들과 대응되는 적어도 일부의 객체들 사이의 깊이 공간에서의 거리가 연관되도록 상기 피처 추출 네트워크를 학습시키는 단계를 포함할 수 있다. https://scholar.dgist.ac.kr/handle/20.500.11750/60421 Title: Influence of scintillation light confinement on depth-of-interaction measurement performance in a single-ended readout PET detector Author(s): Shim, Hyeong Seok; Cho, Min Jeong; He, Wen; Lee, Min Sun; Levin, Craig S.; Lee, Jae Sung Abstract: Continuous depth-of-interaction (cDOI) detectors enable single-ended readout in positron emission tomography (PET) by encoding the interaction depth into the scintillation light distribution. This study presents a comprehensive performance optimization of a cDOI detector based on light distribution tailoring with crossed triangular-shaped reflectors by analyzing the effects of optical geometry based on various DOI decoding algorithms. Two parameters were systematically varied: the degree of light confinement through (1) optical segmentation and (2) crystal pitch adjustment. Five DOI decoding strategies-variance, max/sum ratio, Euclidean-distance classification, Gaussian and modified maximum-likelihood estimation, and artificial neural network (ANN) decoding-were adopted with 8 & times; 8 SiPM readout data. Results show that moderate segmentation (2 & times; 2 configuration) achieved the best DOI precision, yielding a 4.7 mm full width at half maximum (FWHM) and an ANN classification accuracy of 89 %. In the pitch study, 1.5 mm-pitch detector achieved better performance than 3.0 mm-pitch, indicating that increased optical interfaces allows more accurate encoding of the depth-dependent light distribution within the SiPM array. The ANN decoder consistently outperformed in DOI resolution compared to analytical and statistical methods by learning nonlinear spatial correlations among SiPM pixels. These findings highlight the coupled importance of optical geometry and data-driven decoding for achieving high DOI sensitivity in next-generation PET detector designs. 2026-04-30T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60407 Title: Post-Quantum Cryptography Migration on V2X Certificate using KpqC Algorithms Author(s): Seo, Yujin; Kim, Young-Sik Abstract: Connected vehicles utilizing Vehicle-to-Everything (V2X) communication enhance road safety and transportation efficiency, supporting cooperative autonomous driving through real-time interactions. However, increased connectivity raises cyber-attack risks, endangering driver and pedestrian safety. This highlights the urgent need to integrate Post-Quantum Cryptography (PQC) into vehicular communications [6]. In this paper, we implement the Korean PQC digital signature algorithm HAETAE for V2X environments and compare its performance with the NIST PQC signature scheme, ML-DSA which is derived from CRYSTALS-DILITHIUM, and traditional signatures, RSA, and ECC under TLS 1.3 environments. Results indicate that PQC algorithms introduce substantial overhead, whereas traditional algorithms produce smaller certificates. Specifically, HAETAE provides more efficient certificates than ML-DSA [3], minimizing latency impacts in TLS operations. These findings inform the critical balance between enhanced security and certificate size, guiding future post-quantum TLS designs. 2025-07-07T15:00:00Z