讲座题目：A Fast Hyperplane-based Minimum-Volume Enclosing Simplex Algorithm for Blind Hyperspectral Unmixing
主讲人：Prof.Chong-Yung Chi (National Tsing Hua University)
内容摘要：Hyperspectral unmixing (HU) is a crucial signalprocessing procedure to identify the underlying materials (or endmembers)and their corresponding proportions (or abundances)from an observed hyperspectral scene. A well-known blind HUcriterion, advocated by Craig during the early 1990s, considersthe vertices of the minimum-volume enclosing simplex of the datacloud as good endmember estimates, and it has been empiricallyand theoretically found effective even in the scenario of no purepixels. However, such kinds of algorithms may suffer from heavysimplex volume computations in numerical optimization, etc. Inthis talk, without involving any simplex volume computations,by exploiting a convex geometry fact that a simplest simplex of N vertices can be defined by N associated hyperplanes, a fast blind HU algorithm that was recently published is introduced, for which each of the N hyperplanesassociated with the Craig’s simplex of N vertices is constructedfrom N-1 affinely independent data pixels, together with anendmemberidentifiability analysis for its performance support.Without resorting to numerical optimization, the devised algorithmsearches for the N(N-1) active data pixels via simplelinear algebraic computations, accounting for its high computationalefficiency. Monte Carlo simulations and real data experiments areprovided to demonstrate its consistent superior efficacy over some benchmarkCraig-criterion-based algorithms in both computationalefficiency and estimation accuracy.