Increasing bit depth of an image from 4 bit to 8 bit

Hi I have an image of skull in gray format and i first decreased its bit depth from 8 bit to 4 bit,which was easy but now the problem is with recovering the original image from this transformed 4 bit image to 8 bit.....hope u got my point... waiting for the reply..

[Jan]

Please post how you converted it to 2 bits. As far as I can see, some color information is lost by this procedure. Then you cannot recover the original image.

[Muhammad Hamid]

i have used "bitand" fuction for this reason

Image=bitand('originalimage',240)

now it is reduced to 4 bit

Answer 1

[Walter Roberson]

You cannot recover the full depth of an image if you reduced the image to below ceil(log2(N)) bits per pixel where N is the number of unique colors (or shades) in the image ( N = length(unique(IMG(:))) )

[Muhammad Hamid]

sorry i was misaken its 4bit depth not 2bit

[Jan]

@Muhammad: This does not change the reply. If the bit depths is reduced from 8 bit to 4 bit, 4 bits of information are lost.

[Muhammad Hamid]

actually i have done this on 1 image but now i am unable to do on this image because in this image the shades and pixels values are varying after single or double pixel,but in case of what i had done there was a regular pattern so it was easy....

[Walter Roberson]

Yep, that's the sort of thing that happens when you reduce the bit-depth of an image.

Perhaps if you explained more about _why_ you are reducing the bit depth, and what your goals are, we might be able to make suggestions.

[Image Analyst]

"i have done this on 1 image but now i am unable to do on this image" Huh??? Did you do it or not? Are there two images, one that you "did" it on and one that you could not? Bottom line is: Just don't do that! Keep it in 8 bit, or 4 or 8 bytes if using floating point. Don't toss away your data.

[Muhammad Hamid]

@walter:actually it was a task in my DIGITAL IMAGE PROCESSING lab in which first we had to reduce the bit depth and than by some technique we had to recover the image....its not necessary to recover exactly the original...lil bit difference would be acceptable..

[Muhammad Hamid]

@Image Analyst: hahaha yes there were 2 images, sorry for my english..

[Image Analyst]

Why would it work on one and not the other? Just divide by something (16) to get your image into the range 0-15 (a 4 bit range), and then multiply that by 16 to get it into the range 0-255 (an 8 bit range). Explain why you're unable to divide and multiply by 16 on *one* of the images. What error message do you get?