BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Fast Fusion of Multi-band Images: A Powerful Tool for Super-resolu
 tion - Qi Wei (University of Cambridge)
DTSTART:20160331T133000Z
DTEND:20160331T150000Z
UID:TALK65399@talks.cam.ac.uk
CONTACT:Yingzhen Li
DESCRIPTION:Hyperspectral (HS) imaging\, which consists of acquiring a sam
 e scene in several hundreds of\ncontiguous spectral bands (a 3D data cube)
 \, has opened a new range of relevant applications\, such\nas target detec
 tion [Manolakis and Shaw\, 2002]\, classification [C.-I Chang\, 2003] and 
 spectral un-\nmixing [Bioucas-Dias et al.\, 2012]. However\, while HS sens
 ors provide abundant spectral informa-\ntion\, their spatial resolution is
  generally more limited. Thus\, fusing the HS image with other highly\nres
 olved images of the same scene\, such as multispectral (MS) or panchromati
 c (PAN) images is\nan interesting problem\, also known as multi-resolution
  image fusion [Amro et al.\, 2011] (Fig. 1).\nFrom an application point of
  view\, this problem is also important as motivated by recent national\npr
 ograms\, e.g.\, the Japanese next-generation space-borne hyperspectral ima
 ge suite (HISUI)\, which\nfuses co-registered MS and HS images acquired ov
 er the same scene under the same conditions\n[Yokoya and Iwasaki\, 2013]. 
 Bayesian fusion allows for an intuitive interpretation of the fusion proce
 ss\nvia the posterior distribution. Since the fusion problem is usually il
 l-posed\, the Bayesian methodology\noffers a convenient way to regularize 
 the problem by defining appropriate prior distribution for the\nscene of i
 nterest.\n\nIn this work\, a new multi-band image fusion algorithm to enha
 nce the resolution of HS image\nhas been proposed. By exploiting intrinsic
  properties of the blurring and down-sampling matrices\,\na much more effi
 cient fusion method has been developed thanks to a closed-form solution fo
 r the\nSylvester matrix equation associated with maximizing the likelihood
 . The main contribution of this\nfusion scheme is that it gets rid of any 
 simulation-based or optimization-based algorithms which\nare quite time co
 nsuming. Coupled with the alternating direction method of multipliers and 
 the block\ncoordinate descent\, the proposed algorithm can be easily gener
 alized to incorporate different priors or\nhyper-priors for the fusion pro
 blem\, allowing for Bayesian estimators. This method has been applied\nto 
 both the fusion of MS and HS images and to the fusion of PAN and HS images
 . We have tested the\nproposed algorithm in both synthetic data and real d
 ata. Results show that the proposed algorithm\ncompares competitively with
  existing algorithms with the advantage of reducing the computational\ncom
 plexity significantly.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR