Transcript multi-histogram-equalization

Multi-Histogram Equalization Methods
for Contrast Enhancement and Brightness
Preserving
指導教授：萬書言
報告學生：賴薏如
作者：David Menotti, Laurent Najman, Jacques Facon,
and Arnaldo de A. Araújo
IEEE Transactions on Consumer Electronics, Vol. 53, No. 3, AUGUST
2007
Histogram equalization(直方圖均勻化)
利用影像對比統計圖用來決定灰度分佈資
訊，讓我們了解灰度分佈，再進行影像處
理。
 是ㄧ種簡單而有效的影像對比修改技術

Histogram equalization(2)


Change the mean brightness of the image to the
middle level of the gray_level range
強化對比的方式是將長條圖裡每個強度分佈的數
量，依照整個畫面的比例，分配一個對照的新值，
強調的是所謂的均勻分佈概念。
Multi-Histogram equalization
Consists of decomposing the input
image into several sub-images,and then
applying the Histogram equalization to each
one

¥ ： natural numbers，¢ ： integer numbers
is a sub set of
0 ≤x < m , 0 ≤ y < n ,
Image I:
to ¢ L={0,..., L−1}
L is typically 256 ,a point(x,y)∈
, l=I(x,y) is called the level of
the point (x, y) in I .
0 ≤ ls ≤ lf <L, I[ls,lf ]⊆ I,
is a sub set of
, I(x,y) = l
:absolute frequency of the level l in the image I
0 ≤ l ≤ L−1,
0 ≤ls≤l≤lf <L ,
：relative frequency (or the probability) of level l in the
(sub-)image I[ls,lf ]
,0≤ls≤l≤lf ≤L−1
, 0≤ls≤l≤lf ≤L−1
Mean:
Standard deviation:
Shannon's Entropy (熵 )
Minimum Within-Class Variance
MHE (MWCVMHE)
Minimum Middle Level Squared
Error MHE
(MMLSEMHE).


(ls+lf)/2, O[ls,lf ]
middle value of the image
Classical HE method
: uniform histogram of the output image
,ls ≤ l ≤ lf
Other Method
Brightness Bi-HE Method (BBHE)
 Dualistic Sub-Image HE Method (DSIHE)
 Minimum Mean Brightness Error Bi-HE
Method(MMBEBHE)
 Recursive Mean-Separate HE Method
(RMSHE)

TEST RESULTS
TEST RESULTS