• 제목/요약/키워드: Multi-Reference

검색결과 937건 처리시간 0.043초

No-Reference Image Quality Assessment based on Quality Awareness Feature and Multi-task Training

  • Lai, Lijing;Chu, Jun;Leng, Lu
    • Journal of Multimedia Information System
    • /
    • 제9권2호
    • /
    • pp.75-86
    • /
    • 2022
  • The existing image quality assessment (IQA) datasets have a small number of samples. Some methods based on transfer learning or data augmentation cannot make good use of image quality-related features. A No Reference (NR)-IQA method based on multi-task training and quality awareness is proposed. First, single or multiple distortion types and levels are imposed on the original image, and different strategies are used to augment different types of distortion datasets. With the idea of weak supervision, we use the Full Reference (FR)-IQA methods to obtain the pseudo-score label of the generated image. Then, we combine the classification information of the distortion type, level, and the information of the image quality score. The ResNet50 network is trained in the pre-train stage on the augmented dataset to obtain more quality-aware pre-training weights. Finally, the fine-tuning stage training is performed on the target IQA dataset using the quality-aware weights to predicate the final prediction score. Various experiments designed on the synthetic distortions and authentic distortions datasets (LIVE, CSIQ, TID2013, LIVEC, KonIQ-10K) prove that the proposed method can utilize the image quality-related features better than the method using only single-task training. The extracted quality-aware features improve the accuracy of the model.

Variable Structure Model Reference Adaptive Control, for SIMO Systems

  • mohammadi, Ardeshir Karami
    • 제어로봇시스템학회:학술대회논문집
    • /
    • 제어로봇시스템학회 2004년도 ICCAS
    • /
    • pp.1987-1992
    • /
    • 2004
  • A Variable Structure Model Reference Adaptive Controller (VS-MRAC) using state Variables is proposed for single input multi output systems. . The structure of the switching functions is designed based on stability requirements, and global exponential stability is proved. Transient behavior is analyzed using sliding mode control and shows perfect model following at a finite time. The effect of input disturbances on stability and transients is investigated and shows preference to the conventional MRAC schemes with integral adaptation law. Sliding surfaces are independent of system parameters and therefore VS-MRAC is insensitive to system parameter variations. Simulation is presented to clear the theoretical results.

  • PDF

IVERSON언어의 참고어로의 구성(외부매체의 동질화를 위하여) (A Formalism of Iverson Language as a Reference Language for the Organization of Homogeneous Field using many Extermal Media)

  • 김영택
    • 전기의세계
    • /
    • 제24권4호
    • /
    • pp.71-72
    • /
    • 1975
  • A formalism of reference language and homogeneous field is constructed for Iverson Language to organize a virtual file among external media. To execute some data manipulations among these external media files must be organized homogeneously forming multi-dimensional array like single media. This paper shows some organization of reference language to build the virtual file using many external media and some examples of program and hardware organization is presented for the justification of proposals.

  • PDF

Analysis of Reference Inquiries in the Field of Social Science in the Collaborative Reference Service Using the Co-Word Technique

  • 조재인
    • 한국문헌정보학회지
    • /
    • 제49권1호
    • /
    • pp.129-148
    • /
    • 2015
  • This study grasped the true nature of the inquiry domain by analysing the requests for collaborative reference service in the social science field using the co-word technique, and schematized the intellectual structure. First, this study extracted 748 uncontrolled keywords from inquiries for reference in the field of social science. Second, calculated similarity indices between the words on the basis of co-occurrence frequency, and performed not only clustering but also MDS mapping. Third, to grasp the difference in inquiries for reference by period, dividing the period into two parts, and performed comparative analysis. As a result, there formed 5 clusters and "Korea Education" showed an overwhelming size with 40.3% among those clusters. The result of the analysis through the period division showed there were many questions about "Education" during the first half, while a lot of inquiries with focus on "welfare and business information" during the second half.

자체 제작한 근접방사선치료용 다목적 팬텀의 유용성 평가 (Effectiveness of a Custom-made Multi Purpose Brachytherapy Phantom)

  • 장인기;이정진;김완선
    • 대한방사선치료학회지
    • /
    • 제18권2호
    • /
    • pp.119-125
    • /
    • 2006
  • 목 적: 본 연구에서는 근접치료 시 선량기준점에서 선량을 정확히 분석하고자 근접방사선치료용 다목적 팬텀(Multi Purpose Brachytherapy Phantom, MPBP)을 제작하여 그 유용성을 평가하고자 하였다. 대상 및 방법: 자체 제작된 근접방사선치료용 다목적 팬텀(MPBP)에서 다기능삽입구(Multi Function Applicator, MFA)를 이용하여 치료 시와 동일한 조건을 재현한 후 열형광선량계(TLD)를 이용하여 전산화 치료계획 장치에서 계산된 좌표와 동일한 기준점에서 선량을 측정하였다. 측정 대상은 근접방사선 치료환자 중 탄뎀(tandem)과 난형체(ovoid)를 사용한 자궁경부암 환자 4명을 대상으로 환자 당 5회씩 총 20회 A point와 B point 그리고 방광에서 선량을 측정하였다. 결 과: MPBP에서 TLD의 측정 결과는 A Point에서 측정한 환자의 상대오차가 각각 -3.2%, -0.2%, 0%, 3.8%를 보였고, B point에서는 -1.4%, -1.4%, 2.4%, 4.0%를 보였다. 그리고 방광에서는 1.3%, 2%, 5.4%, 7.15%의 상대오차를 나타내었다. 결 론: 자체 제작된 근접치료용 다목적 팬텀(MPBP)은 다기능삽입구(MFA)를 사용하여 치료 시와 동일한 조건에서 선량측정의 재현성을 이룰 수 있었고, A, B point지점과 방광의 선량을 정확히 분석할 수 있었다. 따라서 본 연구에서 제작된 근접방사선치료용 다목적 팬텀은 기준점 선량분석에 매우 유용한 것으로 판단된다.

  • PDF

다중시점 환경에서의 슈퍼픽셀 세그먼테이션 기반 깊이 영상 개선 알고리즘 (Depth Map Correction Algorithm based on Segmentation in Multi-view Systems)

  • 정우경;한종기
    • 방송공학회논문지
    • /
    • 제25권6호
    • /
    • pp.954-964
    • /
    • 2020
  • 실감형 미디어에서 현실감을 느끼게 하는 가장 중요한 요소는 깊이 정보이다. 따라서 고품질의 실감형 미디어를 제작하기 위해서는 고품질의 깊이 정보를 획득하는 것이 필수적이다. 본 논문에서는 고품질의 깊이 정보를 획득하기 위하여 다중 시점 환경에서 깊이 지도를 개선하기 위하여 깊이 지도를 여러 개의 세그먼트로 분할 및 다중 시점간의 관계를 고려하는 알고리즘을 제안한다. 제안된 알고리즘은 슈퍼픽셀 세그먼테이션 기법을 사용하여 기준 시점의 깊이 지도를 여러 세그먼트로 나누고, 각 세그먼트를 인접 시점으로 투영한다. 이후 투영된 세그먼트의 정보를 이용하여 인접 시점의 깊이 지도를 평면 추정을 이용하여 개선한 후, 기준 시점으로 역투영된다. 여러 개의 인접 시점에 대해 이 과정을 반복하여 개선된 인접 시점들의 값들과 기준 시점의 초기 깊이 지도를 가중치 합으로 갱신하여 깊이 지도를 개선한다. 기존 다중 시점 스테레오 비전 알고리즘에 제안된 알고리즘을 적용한 시뮬레이션을 통해 제안된 알고리즘의 결과가 주관적 및 객관적으로 기존 알고리즘을 능가하는 것을 보인다.

Effect of angulation on the 3D trueness of conventional and digital implant impressions for multi-unit restorations

  • Ozay Onoral;Sevcan Kurtulmus-Yilmaz;Dilem Toksoy;Oguz Ozan
    • The Journal of Advanced Prosthodontics
    • /
    • 제15권6호
    • /
    • pp.290-301
    • /
    • 2023
  • PURPOSE. The study aimed to determine the influence of implant angulation on the trueness of multi-unit implant impressions taken through different techniques and strategies. MATERIALS AND METHODS. As reference models, three partially edentulous mandibular models (Model 1: No angulation; Model 2: No angulation for #33, 15-degree distal angulation for #35 and #37; Model 3: No angulation for #33, 25-degree distal angulation for #35 and #37) were created by modifying the angulations of implant analogues. Using a lab scanner, these reference models were scanned. The obtained data were preserved and utilized as virtual references. Three intraoral scanning (IOS) strategies: IOS-Omnicam, ISO-Quadrant, and IOS-Consecutive, as well as two traaditional techniques: splinted open tray (OT) and closed tray (CT), were used to create impressions from each reference model. The best-fit alignment approach was used to sequentially superimpose the reference and test scan data. Computations and statistical analysis of angular (AD), linear (LD), and 3D deviations (RMS) were performed. RESULTS. Model type, impression technique, as well as interaction factor, all demonstrated a significant influence on AD and LD values for all implant locations (P < .05). The Model 1 and SOT techniques displayed the lowest mean AD and LD values across all implant locations. When considering interaction factors, CT-Model 3 and SOT-Model 1 exhibited the highest and lowest mean AD and LD values, respectively. Model type, impression technique, and interaction factor all revealed significant effects on RMS values (P ≤ .001). CT-Model 3 and SOT-Model 1 presented the highest and lowest mean RMS values, respectively. CONCLUSION. Splinted-OT and IOS-Omnicam are recommended for multi-unit implant impressions to enhance trueness, potentially benefiting subsequent manufacturing stages.

Multi-camera System Calibration with Built-in Relative Orientation Constraints (Part 1) Theoretical Principle

  • Lari, Zahra;Habib, Ayman;Mazaheri, Mehdi;Al-Durgham, Kaleel
    • 한국측량학회지
    • /
    • 제32권3호
    • /
    • pp.191-204
    • /
    • 2014
  • In recent years, multi-camera systems have been recognized as an affordable alternative for the collection of 3D spatial data from physical surfaces. The collected data can be applied for different mapping(e.g., mobile mapping and mapping inaccessible locations)or metrology applications (e.g., industrial, biomedical, and architectural). In order to fully exploit the potential accuracy of these systems and ensure successful manipulation of the involved cameras, a careful system calibration should be performed prior to the data collection procedure. The calibration of a multi-camera system is accomplished when the individual cameras are calibrated and the geometric relationships among the different system components are defined. In this paper, a new single-step approach is introduced for the calibration of a multi-camera system (i.e., individual camera calibration and estimation of the lever-arm and boresight angles among the system components). In this approach, one of the cameras is set as the reference camera and the system mounting parameters are defined relative to that reference camera. The proposed approach is easy to implement and computationally efficient. The major advantage of this method, when compared to available multi-camera system calibration approaches, is the flexibility of being applied for either directly or indirectly geo-referenced multi-camera systems. The feasibility of the proposed approach is verified through experimental results using real data collected by a newly-developed indirectly geo-referenced multi-camera system.

공동주택의 에너지 자립을 위한 핵심요소기술의 에너지 성능평가 (Energy Performance Evaluation of Zero Energy Technologies for Zero Energy Multi-House)

  • 윤종호;김병수
    • 한국태양에너지학회 논문집
    • /
    • 제27권3호
    • /
    • pp.161-167
    • /
    • 2007
  • Zero Energy Multi-House(ZeMH) signifies a residential building which can be self sufficient with just new and renewable energy resources without the aid of any existing fossil fuel. For success of ZeMH, various innovative energy technologies Including passive and active systems should be well integrated with a systematic design approach. The first step for ZeMH is definitely to minimize the conventional heating and cooling loads over 50% with major energy conservation measure and passive solar features which are mainly related to building design components such as super-insulation, super window, including infiltration and ventilation issues. The purpose of this study is to analyze the thermal effect of various building design components in the early design of ZeMH. The process of the study is presented in the following. 1) selection reference model for simulation 2) verification of reference model with computer simulation program(ESP-r 9.0). 3) analysis of effect according to insulation-thickness, kinds of windows, rate of infiltration. and The simulation results indicate that almost 50% savings of conventional heating load in multi-house can be achieved with the optimum design of building components such as super insulation, super window, infiltration, ventilation.

계층적 깊이 영상으로 압축된 다시점 비디오에 대한 디지털 워터마크 기술 (Digital Watermarking Technique of Compressed Multi-view Video with Layered Depth Image)

  • 임중희;신종홍;지인호
    • 한국인터넷방송통신학회논문지
    • /
    • 제9권1호
    • /
    • pp.1-9
    • /
    • 2009
  • 본 논문에서는 리프팅 웨이브렛 변환을 이용한 디지털 영상의 워터마크 기술을 제안하였다. 이 워터마크 기술은 쉽게 비디오와 관련된 콘텐트에 대한 워터마크 기술로 확장 될 수 있다. 그래서 깊이 정보를 포함하는 다시점 영상을 효과적으로 압축할 수 있는 계층적 깊이 영상 구조라는 새로운 콘텐트에 제안 기술을 적용하였다. 이 방법은 기준 시점 영상에만 워터마크를 삽입하므로 그 과정이 간단하다는 장점을 가진다. 그리고 계층적 깊이 영상 구조의 특성상 나머지 시점 영상으로 워터마크가 전파되므로 모든 시점 영상에 대하여 저작권을 보호할 수 있다는 장점도 가지고 있다.

  • PDF