• Title/Summary/Keyword: Dense

Search Result 4,031, Processing Time 0.033 seconds

Point cloud removing and rearrangement for reducing bump on 3D mesh (3D Mesh 의 bump 를 감소시키기 위한 Point Cloud 제거 및 재배열 알고리즘)

  • Cha, Sangguk;Han, Jong-Ki
    • Proceedings of the Korean Society of Broadcast Engineers Conference
    • /
    • 2020.07a
    • /
    • pp.266-268
    • /
    • 2020
  • 본 논문에서는 dense point cloud 의 평면영역에서 발생하는 bump 을 줄이기 위한 방법을 제시한다. 이상적인 point cloud 의 평면영역에서 점의 위치의 차이가 균일하다는 특성을 이용하여 점의 위치를 재구성하는 방식을 제시한다. 또한 더 작은 개수의 점으로 물체를 나타낼 수 있으며, 더 작은 잡음이 나타나는 sparse point cloud 의 성질을 고려하여 dense point cloud 의 점의 개수 또한 감소시킨다. 따라서 제안하는 알고리즘을 적용하여 dense point cloud 의 잡음을 감소시키면 평면영역의 bump 감소 및 점 개수의 감소를 통한 데이터 전송 시 더 작은 크기로 보낼 수 있다.

  • PDF

REALM for Open-domain Question Answering of Korean (REALM을 이용한 한국어 오픈도메인 질의 응답)

  • Kan, Dong-Chan;Na, Seung-Hoon;Choi, Yun-Su;Lee, Hye-Woo;Chang, Du-Seong
    • Annual Conference on Human and Language Technology
    • /
    • 2020.10a
    • /
    • pp.192-196
    • /
    • 2020
  • 최근 딥러닝 기술의 발전에 힘입어 오픈 도메인 QA 시스템의 발전은 가속화되고 있다. 특히 IR 시스템(Information Retrieval)과 추출 기반의 기계 독해 모델을 결합한 접근 방식(IRQA)의 경우, 문서와 질문 각각을 연속 벡터로 인코딩하는 IR 시스템(Dense Retrieval)의 연구가 진행되면서 검색 성능이 전통적인 키워드 기반 IR 시스템에 비해 큰 폭으로 상승하였고, 이를 기반으로 오픈 도메인 질의응답의 성능 또한 개선 되었다. 본 논문에서는 경량화 된 BERT 모델을 기반으로 하여 Dense Retrieval 모델 ORQA와 REALM을 사전 학습하고, 한국어 오픈 도메인 QA에서 QA 성능과 검색 성능을 도출한다. 실험 결과, 키워드 기반 IR 시스템 BM25를 기반으로 했던 이전 IRQA 실험결과와 비교하여 더 적은 문서로 더 나은 QA 성능을 보였으며, 검색 결과의 경우, BM25의 성능을 뛰어넘는 결과를 보였다.

  • PDF

Transformer-based dense 3D reconstruction from RGB images (RGB 이미지에서 트랜스포머 기반 고밀도 3D 재구성)

  • Xu, Jiajia;Gao, Rui;Wen, Mingyun;Cho, Kyungeun
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2022.11a
    • /
    • pp.646-647
    • /
    • 2022
  • Multiview stereo (MVS) 3D reconstruction of a scene from images is a fundamental computer vision problem that has been thoroughly researched in recent times. Traditionally, MVS approaches create dense correspondences by constructing regularizations and hand-crafted similarity metrics. Although these techniques have achieved excellent results in the best Lambertian conditions, traditional MVS algorithms still contain a lot of artifacts. Therefore, in this study, we suggest using a transformer network to accelerate the MVS reconstruction. The network is based on a transformer model and can extract dense features with 3D consistency and global context, which are necessary to provide accurate matching for MVS.

State Normalization and Dense Reward Based Reinforcement Learning Method in Basketball Game. (농구 게임에서 상태 정규화 및 Dense 보상 기반 강화 학습 기법)

  • Choi, Taehyeok;Cho, Kyungeun
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2022.11a
    • /
    • pp.475-477
    • /
    • 2022
  • 최근 강화 학습을 적용한 게임 AI 에 대한 연구가 활발히 진행되고 있다. 하지만 대부분 상용게임은 유한 상태 머신(Finite State Machine, FSM)을 이용한 스크립트 기반 AI 를 사용하기 때문에 복잡한 환경의 게임에서 불안정한 상태로 인해 적절한 강화 학습의 수행이 어렵다. 따라서 본 논문에서는 상용 게임 강화 학습 적용을 위하여 상태 정규화 및 Dense 보상 기반 강화 학습 기법을 제안한다. 제안한 기법을 상용 농구 게임에 적용하고 학습된 모델의 성능을 기존 FSM 기반 AI 와 비교를 통해 성능이 약 80% 증가한 결과를 확인하였다.

Photorealistic Real-Time Dense 3D Mesh Mapping for AUV (자율 수중 로봇을 위한 사실적인 실시간 고밀도 3차원 Mesh 지도 작성)

  • Jungwoo Lee;Younggun Cho
    • The Journal of Korea Robotics Society
    • /
    • v.19 no.2
    • /
    • pp.188-195
    • /
    • 2024
  • This paper proposes a photorealistic real-time dense 3D mapping system that utilizes a neural network-based image enhancement method and mesh-based map representation. Due to the characteristics of the underwater environment, where problems such as hazing and low contrast occur, it is hard to apply conventional simultaneous localization and mapping (SLAM) methods. At the same time, the behavior of Autonomous Underwater Vehicle (AUV) is computationally constrained. In this paper, we utilize a neural network-based image enhancement method to improve pose estimation and mapping quality and apply a sliding window-based mesh expansion method to enable lightweight, fast, and photorealistic mapping. To validate our results, we utilize real-world and indoor synthetic datasets. We performed qualitative validation with the real-world dataset and quantitative validation by modeling images from the indoor synthetic dataset as underwater scenes.

ON FUZZY 𝛽-VOLTERRA SPACES

  • V. CHANDIRAN;S. SOUNDARA RAJAN;G. THANGARAJ
    • Journal of applied mathematics & informatics
    • /
    • v.42 no.1
    • /
    • pp.189-197
    • /
    • 2024
  • The purpose of this paper is to introduce and study the new class of spaces called the fuzzy 𝛽-Volterra spaces with the help of fuzzy β-dense and fuzzy 𝛽-G𝛿 sets. Examples are given to illustrate the concept. Some interesting characterizations of the fuzzy 𝛽-Volterra spaces are established in this paper.

Terminals of Antennal Receptor Cells in the Antennal Lobe of the Bunenv. Pieris rapae L.(Insects, Lepidoptera) (배추흰나비 뇌의 촉각엽에 뻗은 촉각지각신경세포의 축색종말)

  • 이봉희;김우갑
    • The Korean Journal of Zoology
    • /
    • v.31 no.4
    • /
    • pp.334-343
    • /
    • 1988
  • Five types of synaPses were differentiated in the antennal lobe or Pieris rapoe. aev are in general axo-dendritic synapses and have symmetrical contacts. The type I synapses contain the uniformly round vesicles of medium size, and consist of the 11% of all synapses in the antennal lobe. The type synapses are filled with the small flat and densely arranged (19%). The type III synapses possess medium-sized round vesicles and dense core ones(24%). The type IV synapses are characterized by such three kinds of vesicles as small round, small flat and dense core vesicles(33%). The type V synapses exhibit the presence of medium-sized round, large round and dense core vesicles(13%). The removal of the left antenna on its proximal portion caused the type IV boutons of the above ave types in ipsilateral antennal lobe to turn into dark or semidark degenerative changes. Therefore, it was concluded that many of the antennal receptor cells projected into the brain terminate in the antennal lobe to form type IV synapses together with the dendrites. 배추회나비 뇌의 촉각엽에 있는 신경세포들은 5가지의 신경연접을 형성하는 것으로 관찰되었다. 이 신경연접들은 모두 축색과 수상돌기간의 연접으로 대칭연접이었다. 제1형 신경연접은 축색종말내에 중등도크기의 단형 연접소포를 가지고 있었고, 촉각엽에 있는 5가지의 신경연접중 11%의 출현빈도수로 관찰되었다. 제2형 신경연접의 축색종말에는 장형과 도형의 2가지 연접소포가 포함되어 있었는데, 모두 소형이었으며 이같은 연접소포들이 축색 종말내에서 조밀하게 관찰되었다. 제2형 신경연접의 출현 빈도수는 19%이었다. 제3형 신경연접은 축색 종말내에 중등도크기의 도형 연접소포와 DENSE CORE VESICLE을 포함하고 있었고 그 출현빈도수는 24%이었다. 제4형 신경연접은 축색총말내에 소형의 도형, 소형의 장형 연접소포 및 DENSE CORE VESICLE의 3가지 연접소포를 가지고 있었고 출현빈도수는 촉각엽에서 가장 큰 33%이었다. 제5형 신경연접은 축색종말내에 중등도크기의 원형, 대형의 원형연접소포 및 DENSE CORE VESICLE을 포함하였고 13%의 출현빈도수로 관찰되었다. 배추횐나비의 촉각에 있는 지각신경세포가 뇌의 촉각엽으로 뻗어 들어가 위의 5가지 신경연접중 어느 형을 형성하는지를 관찰하기 위하여 좌측 촉각의 기부를 제거하여 지각신경세포를 절단하였는데 그 결과, 좌측 촉각엽에서 제4형의 신경연접이 퇴행성 변화를 나타내었다. 그러므로 촉각의 지각신경세포는 뇌의 같은 족 촉각엽에 뻗어와 제4형 신경연접을 형성한다고 결론되었다.

  • PDF

A Case of Membranoproliferative Glomerulonephritis Type II (Dense Deposit Disease, DDD) (막성증식성 사구체신염 제 II형 (Dense Deposit Disease, DDD) 1례)

  • Kwon Hae Sik;Oh Seung-Jin;Lee Young-Mock;Kim Ji Hong;Kim Pyung-Kil;Kang Hae Youn;Jeong Hyeon Joo;Choi In Joon
    • Childhood Kidney Diseases
    • /
    • v.5 no.2
    • /
    • pp.188-195
    • /
    • 2001
  • Type II membranoproliferative glomerulonephritis (Dense deposit disease) is an acquired primary glomerular disease characterized by electron microscopic evidence of a continuous dense membrane deposition replacing the lamina densa. It is a subtype of idiopathic membra- noproliferative glomerulonephritis, and was described as a separate entity by Berger and Galle in 1963. It frequently occurs in older chilren and young adults and the clinical course is variable, but is generally progressive. The presenting feature is nephrotic syndrome in many patients, and proteinuria and hematuria are also seen frequently. The purpose of this paper is to present a case of DDD (Dense deposit disease) from a 10 year old boy who was diagnosed as a acute poststreptococcal glomurulonephritis with protenuria, hematuria, and facial edema by renal biopsy 4 years ago. (J, Korean Soc Pediatr Nephrol 2001 ; 5 : 188-95)

  • PDF

TRAO Multi-beam Legacy Survey of Nearby Filamentary Molecular Clouds : Progress Report

  • Kim, ShinYoung;Chung, Eun Jung;Lee, Chang Won;Myers, Philip C.;Caselli, Paola;Tafalla, Mario;Kim, Gwanjeong;Kim, Miryang;Soam, Archana;Gophinathan, Maheswar;Liu, Tie;Kim, Kyounghee;Kwon, Woojin;Kim, Jongsoo
    • The Bulletin of The Korean Astronomical Society
    • /
    • v.42 no.1
    • /
    • pp.32.1-32.1
    • /
    • 2017
  • To dynamically and chemically understand how filaments, dense cores, and stars form under different environments, we are conducting a systematic mapping survey of nearby molecular clouds using the TRAO 14 m telescope with high ($N_2H^+$ 1-0, $HCO^+$ 1-0, SO 32-21, and $NH_2D$ v=1-0) and low ($^{13}CO$ 1-0, $C^{18}O$ 1-0) density tracers. The goals of this survey are to obtain the velocity distribution of low dense filaments and their dense cores for the study of their origin of the formation, to understand whether the dense cores form from any radial accretion or inward motions toward dense cores from their surrounding filaments, and to study the chemical differentiation of the filaments and the dense cores. Until Feb. 2017, the real OTF observation time is 460 hours. We have almost completed mapping observation with four molecular lines ($^{13}CO$ 1-0, $C^{18}O$ 1-0, $N_2H^+$ 1-0, and $HCO^+$ 1-0) on the five regions of molecular clouds (L1251 of Cepheus, Perseus west, Polaris south, BISTRO region of Serpense, California, and Orion B). The maps of a total area of $7.38deg^2$ for both $^{13}CO$ and $C^{18}O$ lines and $2.19deg^2$ for both $N_2H^+$ and $HCO^+$ lines were obtained. All OTF data were regridded to a cell size of 22 by 22 arcseconds. The $^{13}CO$ and $C^{18}O$ data show the RMS noise level of about 0.22 K and $N_2H^+$ and $HCO^+$ data show about 0.14 K at the velocity resolution of 0.06 km/s. Additional observations will be made on some regions that have not reached the noise level for analysis. We are refining the process for a massive amount of data and the data reduction and analysis are underway. This presentation introduces the overall progress from observations to data processing and the initial analysis results to date.

  • PDF

A Study on the Digital Mammography for Breast Cancer Patients (유방암 환자의 Digital Mammography에 관한 연구)

  • Lim, Cheong-Hwan;Lee, Sang-Ho;Jung, Hong-Ryang;Mo, Eun-Hui
    • Journal of the Korean Society of Radiology
    • /
    • v.6 no.1
    • /
    • pp.63-71
    • /
    • 2012
  • This study aimed to evaluate the accuracy of breast cancer diagnosis of digital mammography which is in the highest interest of breast imaging test, and to investigate the characteristics of breast cancer patients. For this purpose, 57 breast cancer patients who underwent breast imaging test were examined between May 2010 and June 2011. The average age of the breast cancer patients was 50.8 years old, and the most frequently occurring location was the upper outer quadrant (UOQ), accounting for 33.3%. By age, the highest occurrence rate of breast cancer was the age group of 40~49, accounting for 42.1%. As for the breast composition of the breast cancer patients, fatty breast accounted for 31.6% (18/57) and dense breast for 68.4% (39/57), indicating that nearly 70% of the breast cancer patients have dense breast. It was found that the detection rate of breast cancer was the highest (45.3%) when both microcalcification and mass are simultaneously present in the radiographic lesion of the breast imaging. In dense breast, the mass without microcalcification was lower in detection rate than fatty breast. Accordingly, the mass is the cause of raising the false negative rate in dense breast. The findings show that the false negative rate of digital mammography was 7.0% and the sensitivity 93.0%. Also, the false negative rate of dense breast was 12.8%, and the sensitivity 87.2%, indicating that the sensitivity to breast cancer in this study was higher than the dense breast of previously reported screen film mammography.