Courses
Sections
Projects
Wiki
News
Forums
Surveys
Mailing Lists| Author | Message |
|---|---|
|
yogesh
Posts: 24
|
Posted 20:25 May 13, 2011 |
Jpg comp ration chromosome sub sampling 4:2:0
Cr and cb value are we take Image size 280 X 256 so cr and cb are [144][128] and average of 4 pixel 144X128 280X256 Cr[0,0] = Avg ([0,0]+[0,1]+[1,0][1,1]) | | Cr[144][128] Inver side How to make original size array Cr[144][128]--->nCr[280][256] ? |
|
cbort
Posts: 95
|
Posted 20:33 May 13, 2011 |
You will have to create an array that is the normal size and fill it with cb // cr using each value to fill 4 pixels worth so for the inverse you have
cr[0][0] - > newcr[0][0] cr[0][0] -> newcr[1][0] cr[0][0] -> newcr[0][1] cr[0][0] -> newcr[1][1] cr[0][1] -> newcr[0][2] cr[0][1] -> newcr[0][3] cr[0][1] -> newcr[1][2] cr[0][1] -> newcr[1][3] .... essentially you want cr[i][j] -> cr[i*2][crj*2] cr[i][j] -> cr[i*2+1][crj*2] cr[i][j] -> cr[i*2][crj*2+1] cr[i][j] -> cr[i*2+1][crj*2+1] |
|
yogesh
Posts: 24
|
Posted 20:54 May 13, 2011 |
i do same thing but n=5 i got big dot in my image what wrong with me i did not known ?
|
|
cbort
Posts: 95
|
Posted 20:59 May 13, 2011 |
5 is supposed to look bad....
Professor Kang posted samples on the class site that you can compare to... do your lower level quanitzation (n = [0,2]) look similar too? If not I think it is fine |
