DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kweon, In So | - |
dc.contributor.advisor | 권인소 | - |
dc.contributor.author | Park, Jinsun | - |
dc.contributor.author | 박진선 | - |
dc.date.accessioned | 2017-03-29T02:36:50Z | - |
dc.date.available | 2017-03-29T02:36:50Z | - |
dc.date.issued | 2016 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=649608&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/221667 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 전기및전자공학부, 2016.2 ,[v, 39 p. :] | - |
dc.description.abstract | In this thesis, we introduce robust and synergetic hand-crafted features and a simple but efficient deep feature from a convolutional neural network (CNN) architecture for defocus estimation. This thesis systematically analyzes the effectiveness of different features, and shows how each feature can compensate for weaknesses of other features when they are concatenated. For a full defocus map estimation, we extract image patches on strong edges sparsely, then we use them for the deep and hand-crafted features extraction. In order to reduce patch scale dependency, we also propose multi-scale patch extraction strategy. A sparse defocus map is generated using a neural network classifier followed by a probability-joint bilateral filter. The final defocus map is obtained from the sparse defocus map with a guidance of an edge preserving filtered input image. Experimental results show that our algorithm is superior to state-of-the-art algorithms in defocus estimation. Our work can be used for applications including segmentation, image synthesizing, blur magnification, all-in-focus image generation, and 3-D estimation. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Defocus Estimation | - |
dc.subject | Deep Feature | - |
dc.subject | Hand-crafted Features | - |
dc.subject | Image Processing | - |
dc.subject | Computer Vision | - |
dc.subject | 흐림 추정 | - |
dc.subject | 깊은 특징 | - |
dc.subject | 수제 특징 | - |
dc.subject | 영상 처리 | - |
dc.subject | 컴퓨터 비젼 | - |
dc.title | (A) unified approach of deep and hand-crafted features for defocus estimation | - |
dc.title.alternative | 깊은 특징과 수제 특징을 이용한 흐림 추정의 통합적 접근 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :전기및전자공학부, | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.