matlab中的conv2卷积运算的三种形式

thinkfunny

matlab中的conv2是用于对二维数据进行卷积运算，有三个参数可供选择，下面是help content of conv2
2 Z' z2 n) k* h7 w, i
: J1 H9 n7 \# S/ ^conv2 Two dimensional convolution.
7 ^0 Y; d+ P2 `+ I7 n- p! c
' O! ^& O* d* C0 P  M0 M    C = conv2(A, B) peRForms the 2-D convolution of matrices A and B.
2 _2 D2 v. A6 i! T# L+ q" o    If [ma,na] = size(A), [mb,nb] = size(B), and [mc,nc] = size(C), then
7 O+ ]4 i3 S3 l    mc = max([ma+mb-1,ma,mb]) and nc = max([na+nb-1,na,nb]).
" j( ]$ o1 V1 ^2 g0 ?0 R* g
8 l; d3 u9 N" K    C = conv2(H1, H2, A) first convolves each column of A with the vector
  l$ f4 S. I& e/ R# u! v( d    H1 and then convolves each row of the result with the vector H2.  If
    n1 = length(H1), n2 = length(H2), and [mc,nc] = size(C) then
/ x7 {* z6 `, M% o7 G+ l$ x6 k0 g    mc = max([ma+n1-1,ma,n1]) and nc = max([na+n2-1,na,n2]).
    conv2(H1, H2, A) is equivalent to conv2(H1(：)*H2(：).', A) up to
# H; @2 y( i& w! p    round-off.
" v7 i( J2 p3 @  h* B0 V6 h
    C = conv2(..., SHAPE) returns a subsection of the 2-D
+ M; \2 d& v- g; e0 g    convolution with size specified by SHAPE:
      'full'  - (default) returns the full 2-D convolution,
: A2 P5 C& k1 r      'same'  - returns the central part of the convolution
                that is the same size as A.
/ {% g& G: {2 k      'valid' - returns only those parts of the convolution
                that are computed without the zero-padded edges.
: z$ T. Z1 p  t" S                size(C) = max([ma-max(0,mb-1),na-max(0,nb-1)],0).
; w! s( i* }3 D8 I% O

关于full, same以及valid三种参数的区别，如下面的实例所示：

full
5 ~% h& D( {+ {& E0 P+ G0 _. H- ?



6 [# l/ u7 [' W$ ^* w$ x- n) gsame
* D; m- \; R8 ]. `$ ]& f

$ y6 ^( \7 H, a  h" `1 t8 g& ?+ `) S
2 c% L! K9 x2 I
/ l( U+ T: V) E2 X, w9 svalid
" c- K& I2 u# t1 f) K

/ c0 u$ R. F+ q& O3 E
1 Z# h% C) f, v/ u( T
) d( K( C3 _* _5 r( H/ p. j

yin123

三种形式的matlab中的conv2卷积运算