Local Enhancement 1 Local Enhancement • Local Enhancement • HW2 due Oct 9 • Required Reading: Sections 3.1 – 3.3 • Project groups: need to finalize soon (tomorrow?) Local Enhancement 2 Local enhancement Sometimes Local Enhancement is Preferred. Malab: BlkProc operation for block processing. Left: original “tire” image.Local Enhancement 3 Histogram equalized Local Enhancement 4 Local histogram equalized F=@ histeq; I=imread(‘tire.tif’); J=blkproc(I,[20 20], F);Local Enhancement 5 Fig 3.23: Another example Local Enhancement 6 Local Contrast Enhancement • Enhancing local contrast g (x,y) = A( x,y ) [ f (x,y) - m (x,y) ] + m (x,y) A (x,y) = k M / σ(x,y) 0 < k < 1 M : Global mean m (x,y) , σ (x,y) : Local mean and standard dev. Areas with low contrast Larger gain A (x,y)Local Enhancement 7 Fig 3.27 Figure numbers refer to the second edition Local Enhancement 8 Fig 3.25 Figure numbers refer to the second editionLocal Enhancement 9 Fig 3.26 Figure numbers refer to the second edition Local Enhancement 10 Image Subtraction g (x,y) = f (x,y) - h (x,y) h(x,y)—a low pass filtered version of f(x,y). • Application in medical imaging --“mask mode radiography” • H(x,y) is the mask, e.g., an X-ray image of part of a body; f(x,y) –incoming image after injecting a contrast medium.Local Enhancement 11 Subtraction: an example Figure numbers refer to the second edition Local Enhancement 12 Fig 3.28: mask mode radiography Figure numbers refer to the second
View Full Document