Berkeley EE225B video I note

////////////////////////////////////////////  


17:26 2013-10-14 


start Berkeley EE225B Digital Image Processing 


 


17:34 2013-10-14 


image processing & video processing 


 


17:56 2013-10-14 


image enhancement 


 


18:28 2013-10-14 


ICIP == International Conference for Image Processing 


 


20:03 2013-10-14 


MPEG == Moving Picture Expert Group 


11:26 2013-11-15 


start berkely EE225B, digital image processing 


 


11:29 2013-11-15 


tomography 


 


11:38 2013-11-15 


this picture show an image of a gamma ray, 


 


it goes through flesh easily, but not bones 


 


11:40 2013-11-15 


cygnus loop??? 


 


11:42 2013-11-15 


x rays: medical imaging 


 


11:44 2013-11-15 


What's the primary application of UV imaging? 


 


11:54 2013-11-15 


UV == Ultra Violet 


 


11:57 2013-11-15 


we're now moving from visible to infrared 


 


11:58 2013-11-15 


MultiSpectrum Imaging 


 


12:01 2013-11-15 


this is the map of continent of America 


 


12:03 2013-11-15 


here is a classic inspection machines 


 


12:05 2013-11-15 


ASIC == Application-Specific IC 


 


12:08 2013-11-15 


I'm taling about microwave band now! 


 


12:09 2013-11-15 



SAR == Synthetic Aperture Radar 


 


12:12 2013-11-15 


MRI == Meganetic Radioactive Imaging 


 


12:14 2013-11-15 


E/M == ElectroMagnetic 


 


12:18 2013-11-15 


thermal failure 


 


12:20 2013-11-15 


multidimensional signal processing 


 


12:21 2013-11-15 


M-D signal == MultiDimensional Signal 


 


12:21 2013-11-15 


because females have higher frequency in 


 


their voice & speech 


 


12:26 2013-11-15 


M-D signals are often described by 


 


PDE(Partial Differential Equations) 


 


12:27 2013-11-15 


PDE == Partial Differential Equations 


 


12:28 2013-11-15 


fundamental theorem of algebra??? 


 


does not hold in M-D 


 


12:31 2013-11-15 


How can you factor a polynomial? 


 


12:32 2013-11-15 


IIR filter 


 


IIR == Infinite Impulse Response 


 


12:35 2013-11-15 



impulse response 


 


12:35 2013-11-15 


FIR == Finite Impulse Response 


 


12:36 2013-11-15 


FIR filter has no pole, so it's stable! 


 


12:36 2013-11-15 


BIBO == Bounded Input Bounded Output 


 


12:36 2013-11-15 


BIBO stability 


 


12:37 2013-11-15 


there are two kinds of IIR filters? 


 


1. IIR filter with a rational transfer function 


 


2. IIR filter without a rational transfer function 


 


12:39 2013-11-15 


IIR == Infinite Impulse Response 


 


12:39 2013-11-15 


IIR filter with rational transfer function: 


 


H(z) = P(z) / Q(z) 


 


12:42 2013-11-15 


ratio of polynomial functions 


 


12:42 2013-11-15 


we only worry about IIR filters that have 


 


a rational transfer function! 


 


12:43 2013-11-15 


IIR filter implies feedback! 


 


12:51 2013-11-15 


linear phase filters 


 


12:51 2013-11-15 



in contrast to FIR filters, it's very hard 


 


to get linear phase in IIR filters 


 


13:05 2013-11-15 


unit sequence 


 


13:06 2013-11-15 


notation for 2-D sequences 


 


13:06 2013-11-15 


line impulse 


 


13:10 2013-11-15 


step sequence 


 


13:14 2013-11-15 


linear system 


 


13:14 2013-11-15 


shifting invariance 


 


13:16 2013-11-15 


LSI == Linear Shifting Invariance 


 


13:16 2013-11-15 


LSI system can be completely characterized by 


 


impulse reponse! 


 


13:19 2013-11-15 


linearize the nonlinear system! 


 


13:20 2013-11-15 


from the historical point of view, why 


 


we stick with LSI sytem? 


 


because LSI sytem has such a nice feature 


 


that it can be complelely described by 


 


impulse response function! 


 



thus y = x * h! 


 


13:23 2013-11-15 


LTI == Linear Time-Invariant 


 


13:25 2013-11-15 


two dimensional convolution 


 


13:30 2013-11-15 


What is a separable sequence? 


 


a separable sequence is one that can be 


 


written as the product of x1(n)x2(n) 


 


13:35 2013-11-15 


if the system is separable,... 


 


13:36 2013-11-15 


separable LSI system 


 


13:45 2013-11-15 


2D convolution??? 


 


13:47 2013-11-15 


separable LSI system 


 


nonseparable LSI system 


 


13:50 2013-11-15 


stability 


 


13:50 2013-11-15 


BIBO == Bounded Input Bounded Output 


 


13:51 2013-11-15 


finite dynamic range 


 


13:53 2013-11-15 


marginal stable system 


 


13:56 2013-11-15 


you have to ensure that the impulse response 


 



is absolutely summable! 


 


13:56 2013-11-15 


FIR always stable! 


 


13:57 2013-11-15 


IIR some stable, some unstable! 


 


13:57 2013-11-15 


blow up! 


---------------------------------------- 


 


sequence which can be written as the 


 


product of x1(n1) * x2(n2)! 


 


13:42 2013-11-16 


we have an n * n input signal 


 


13:43 2013-11-16 


essentially what we're doing here is 


 


that... 


------------------------------------------