BISPL- BioImaging, Signal Processing & Learning Lab @ KAIST AI
  • BISPL News
  • Professor
  • Members
  • Research
  • Publications
  • Github
  • Software
    • NIRS-SPM
    • k-t FOCUSS
    • Compressive MUSIC
    • Super-resolution Microscopy
    • Patch Low Rank MRI
    • Sparse SPM
    • ALOHA Inpainting
    • ALOHA for MR Recon
    • MR Artifact Removal using Robust ALOHA
    • MR Ghost Artifact correction using ALOHA
  • BISPL Hall of Fame
  • Recent Talks
  • COVID-19 data link
  • BISPL Snapshots
  • On the World Press
  • Intranet

NIRS-SPM

k-t FOCUSS

Compressive MUSIC

Super-resolution Microscopy

Patch Low Rank MRI

Sparse SPM

ALOHA Inpainting

ALOHA for MRI

Annihilating Filter-based Low-Rank Hankel Matrix (ALOHA) for Image Inpainting

  1. Kyong Hwan Jin and Jong Chul Ye, "Annihilating filter based low rank Hankel matrix approach for image inpainting",  IEEE Trans. Image Processing, vol. 24, no. 11, pp 3498-3511, 2015.
Picture
   Abstract—In this paper, we propose a patch-based image inpainting method using a low-rank Hankel structured matrix completion approach. The proposed method exploits the annihilation property between a shift-invariant filter and image data observed in many existing inpainting algorithms. In particular, by exploiting the commutative property of the convolution, the annihilation property results in a low-rank block Hankel structure data matrix, and the image inpainting problem becomes a low-rank structured matrix completion problem. The block Hankel structured matrices are obtained patch-by-patch to adapt to the local changes in the image statistics. To solve the structured low-rank matrix completion problem, we employ an alternating direction method of multipliers with factorization matrix initialization using the low-rank matrix fitting algorithm. As a side product of the matrix factorization, locally adaptive
dictionaries can be also easily constructed. Despite the simplicity of the algorithm, the experimental results using irregularly subsampled images as well as various images with globally missing patterns showed that the proposed method outperforms existing state-of-the-art image inpainting methods.
Picture
Benchmark comparison with state-of-the art inpainting algorithms 



    Release note:  version 1.0, Fe. 20, 2016

Download
ABOUT US
Our research activities are primarily focused on the signal processing and machine learning  for high-resolution high-sensitivity image reconstruction from real world bio-medical imaging systems. Such problems pose interesting challenges that often lead to investigations of fundamental problems in various branches of physics, mathematics, signal processing, biology, and medicine. While most of the biomedical imaging researchers are interested in addressing this problem using off-the-self tools from signal processing, machine learning, statistics, and optimization and combining their domain-specific knowledge, our approaches are unique in the sense that I believe that actual bio-medical imaging applications are a source of endless inspiration for new mathematical theories and we are eager to solve both specific applications and application-inspired fundamental theoretical problems. 
CONTACT US
Bio Imaging. Signal Processing & Learning
Graduate School of AI
KAIST
291 Daehak-ro, Yuseong-gu
Daejeon 305-701, Korea

Copyright (c) 2014, BISPL
All Rights Reserved.

Picture