Hello, I am trying to find out how the Stiles and Stile_origin parameters of kdu_compress.exe are working. I would like to compress an image with multiple...
J. Andert
Juergen.Andert@...
Jul 1, 2002 12:07 pm
268
Sorry, there are no such functions provided at this point which make this easy. The problem is that to really calculate the bit-rate you need to read in or...
Hi all, I'm new to the kakadu jpeg2000 codec and also the jpeg2000 algorithm. Do you think kakadu can achieve zero distortion in the region-of-interest at a...
Surene Chan
surene.chan@...
Jul 2, 2002 5:55 am
270
When I code the images using the kakadu with different progressions LRCP and RLCP and see the same in kdu_show there is no difference. Progressive order by...
When I compress the images using the kakadu using the region of interest mask. When I see the images using the kdu_show the region of interest seems to grow. ...
Dear sirs, I try to compress 16 bit medical images to raw file and decompress back to PGM file. When I choose reversible(lossless), everything works fine. if I...
Sky, You probably got the endian format wrong. kdu_compress expects big endian. If you chose the wrong endian format then for 16-bit shorts you will start cut...
Juergen, Stiles tells kakadu at what pixel position the first (upper left) tile starts. Stile_origin is the size of each tile. So you can have something like: ...
I rather think it's kdu_show what does not differate between LRCP and RLCP. I hope I am wrong, but I have used other encoders to produce both orders by quality...
... Hi Mark and Atti... I think Kakadu-show supports every progression-order. The point is, LRCP and RLCP start both with the first layer of the 0-Resolution....
... Hi atti... Yes, it is. To decode a RoI losslessly you have to consider all coefficients affecting the RoI. Due to the multi-res.-representation in J2k,...
Hello, As I was comparing lossless JPEG2000 compression with lossless JPEG, I observed that JPEG2000 is pretty slow as compared to JPEG, and besides there is...
Hi everybody, I have a question about the PLM Marker : does Kakadu insert it in the .j2c file automatically when we compress an image? Can anybody help me ? ...
Hi, I have been compiling and testing Version 2.2 code that came with the book. When trying to compress an 8-bit DIB bitmap, kdu_compress complains about...
if we use Enumerated Colorspace specification method (for 3 component images)then which are the color transforms to be used? if we use restricted ICC profile...
Surya and All, If you do not need the scalability of j2k and only use lossless compression or low compression ratios then IMHO there are better solutions out...
I'm not sure I fully understand your question. The specification of colorspace (using either the Enumerated or Restricted ICC method) in a JP2 file is...
Rob, Thanks for your reply, and sorry to respond so late because I'm out for vocation last week. Endian format is not the reason. In fact, what I'm doing now...
Sky, It's hard to judge what is going wrong and what you understand under big distortion. Here just some ideas: 1) Obviously you are going to have some errors...
Hello, Has anyone conducted comparisons between the compressing on either of the colorspaces. My understanding is tha the default setting on Kakadu is to make...
Gerardo, You do not necessary get better results with YCC then RGB. Also the PSNR is just a mathematical number that does not say much about the actual...
Rob and Gerardo, Generally, for natural images, you should get somewhat better results using YCC for compression than RGB. The colorspace transform is used for...
Thanks a lot for the reply. As I understand from what you said, from an ENCODER point of view, is the only difference between co-existing EnumSpace &...
Hi All, I've not brought the source-code yet, but have evaluated the demo tools and I'm very impressed with them, especially the interactive nature of the JPIK...
In general, if your input image data is stored in an RGB colorspace, you should use the component transformation. The selection of the RCT or the ICT is ...
... Yes, this is correct. ... It comes in on decode when your application interprets the decoded data. For example, if you have an image that specifies its...
John, I am not sure I agree with you here. I think the main reason ycc is used is due to the nature of the human eye. I agree with you that the transform does ...
Hi Robert, Point taken, I should have phrased my comment as: "The colorspace transform is used for compression not only due to the human eye, but also because...
Deng, You also need to shift the sample values for fixed point representation. So you have to check if you use absolute values. If so just leave it as it is. ...
Hi all, Which category is appropriate for JPIK, tier1 or tier3? If I use kdu_show with kdu_server, it seems that it belongs to tier3 as it supports any...
