http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
영상 내 영역 별 지식을 활용한 비지도 학습 기반 의복 변환 상황에서의 사람 재식별 기술
오민영(Minyoung Oh),심재영(Jae-Young Sim) 대한전자공학회 2024 대한전자공학회 학술대회 Vol.2024 No.6
Most of person re-identification methods only consider the datasets where the clothes do not change for a same identity of each person, which lacks real-world applicability. On the other hand, clothes-changing person re-identification considers the scenario that people change their clothes. However, collecting clothes-changing person re-identification datasets requires significant labor cost due to additional annotations such as clothes identities. In this paper, we propose an unsupervised clothes-changing person reidentification method to alleviate the labeling burden while utilizing region-specific knowledge of person image. Specifically, we improve the clustering performance based on the observation that the upper part features of the person images with a same identity tend to be consistent regardless of the clothes. Experimental results show that the proposed method significantly improves the performance of the clothes changing person reidentification.
사람-의상 군집화 알고리즘 기반 의상 변경 상황에서의 비지도 도메인 적응형 사람 재식별 기술
김윤한(Yoon-Han Kim),오민영(Minyoung Oh),홍승빈(Seung-Bin Hong),심재영(Jae-Young Sim) 대한전자공학회 2024 대한전자공학회 학술대회 Vol.2024 No.6
Most of existing person re-identification (ReID) methods only consider short-term scenarios such that all the instances with the same person identity wear consistent clothes in dataset. Recently, there has been a lot of researches in the field of clothes-changing ReID (CC-ReID) to solve a more realistic problem, assuming that the persons change their clothes over time. However, it is much harder to collect labeled datasets for CC-ReID. To deal with this problem, we propose a novel unsupervised domain adaptive setting for CC-ReID, which is based on the clustering methods. Meanwhile, existing clustering algorithms depend on the person identity as well as the characteristics of data such as the clothes, light conditions, camera settings, etc. These may cause poor clustering performance in CC-ReID. Therefore, we propose a Cross Clustering Module (CCM) that performs the pseudo labeling by considering not only the person features but also the clothing features. Experimental results verify the proposed method improves the performance of CC-ReID by a large margin on LTCC dataset.
Dienes의 수학학습이론에 따른 사다리꼴의 넓이 학습에서 학생들이 구성한 예 공간 분석
오민영 ( Oh¸ Min Young ),김남균 ( Kim¸ Nam Gyun ) 한국수학교육학회 2021 初等 數學敎育 Vol.24 No.4
The area of a trapezoid is an important concept to develop mathematical thinking and competency, but many students tend to understand the formula for the area of a trapezoid instrumentally. A clue to solving these problems could be found in Dienes’ theory of learning mathematics and Watson and Mason’ concept of example spaces. The purpose of this study is to obtain implications for the teaching and learning of the area of the trapezoid. This study analyzed the example spaces constructed by students in learning the area of a trapezoid based on Dienes’ theory of learning mathematics. As a result of the analysis, the example spaces for each stage of math learning constructed by the students were a trapezoidal variation example spaces in the play stage, a common representation example spaces in the comparison-representation stage, and a trapezoidal area formula example spaces in the symbolization-formalization stage. The type, generation, extent, and relevance of examples constituting example spaces were analyzed, and the structure of the example spaces was presented as a map. This study also analyzed general examples, special examples, conventional examples of example spaces, and discussed how to utilize examples and example spaces in teaching and learning the area of a trapezoid. Through this study, it was found that it is appropriate to apply Dienes’ theory of learning mathematics to learning the are of a trapezoid, and this study can be a model for learning the area of the trapezoid.