컴퓨터비전, 머신러닝, 딥러닝을 이용한 의료영상분석 10-7. Super-Resolution via dictionary learning
컴퓨터비전, 머신러닝, 딥러닝을 이용한 의료영상분석 10-7. Super-Resolution via dictionary learning
본 내용은 Edwith의 컴퓨터비전, 머신러닝, 딥러닝을 이용한 의료영상분석을 요약 정리한 내용으로 DGIST 박상현 교수님과 Edwith, STAR-MOOC에 그 저작권이 있음을 미리 공지합니다.
URL : https://www.edwith.org/medical-20200327/lecture/63182
Super-Resolution
https://ieeexplore.ieee.org/document/5466111
Image Super-Resolution Via Sparse Representation
This paper presents a new approach to single-image superresolution, based upon sparse signal representation. Research on image statistics suggests that image patches can be well-represented as a sparse linear combination of elements from an appropriately c
ieeexplore.ieee.org
Y : observation
X : SR
$D_l$ : low resolution dictionary
$D_h$ : high resolution dictionary
$min_{\alpha} \left| F(Y) - F(D_l \alpha) \right|_2^2 + \lambda \left| \alpha \right|$
F( ) : Feature
$X = D_h \alpha$
$\alpha$를 찾고 HR Dictionary와 곱을 진행함
patch 단위 processing시 overlapping
