I=imread('lena.jpg');
imwrite(I,'new_lena.jpg','Quality',100);
I1=imread('new_lena.png');
As per imwrite documentation ('Quality',100) pair means no compression but I get a psnr around 59. I expect psnr value = inf. Can anyone explain why and if I am doing something wrong either writing or reading the image?
You are doing all things on the right way. This explanation of Quality criteria is from Mathwork site:
Quality of the JPEG-compressed file, specified as the comma-separated pair consisting of 'Quality' and a scalar in the range [0,100], where 0 is lower quality and higher compression and 100 is the high quality and the less compression.
That means the number 100 make the less compression rate so the result image will be compressed according to the lossy JPEG compression algorithm which means that PSNR rate will be in the range that you got.
You are doing all things on the right way. This explanation of Quality criteria is from Mathwork site:
Quality of the JPEG-compressed file, specified as the comma-separated pair consisting of 'Quality' and a scalar in the range [0,100], where 0 is lower quality and higher compression and 100 is the high quality and the less compression.
That means the number 100 make the less compression rate so the result image will be compressed according to the lossy JPEG compression algorithm which means that PSNR rate will be in the range that you got.
Are you sure it is I1=imread('new_lena.png'); and not I1=imread('new_lena.jpg'); ?
Your question does not really relate to MATLAB with imwrite function. In fact, you want to ask about the compression rate = 1 (100%), will information be lost?
So the answer is https://stackoverflow.com/questions/7982409/is-jpeg-lossless-when-quality-is-set-to-100:
---------------------------------------------------
As correctly answered above, using a "typical" JPEG encoder at quality 100 does not give you lossless compression. Lossless JPEG encoding exists, but it's different in nature and seldom used.
I'm just posting to say why quality 100 does not mean lossless.
In JPEG compression information is mostly lost during the DCT coefficient quantization step (8-by-8 coefficient blocks are divided by a 8-by-8 quantization table, so they become smaller --> 'more compressible'). When you set JPEG quality to 100, no real quantization takes place (because the quantization table will be all 1s, at least with standard IJG-JPEG tables), so in fact you don't lose information here..
However, there are mainly two factors leading to information loss even when no quantization takes place:
Please take into account that he compares with a previously already JPEG encoded file. What you say is true, but it was already done when the first encoding was done.
Robert Alexandru Dobre If he want to compare by PSNR/SSIM, he should use same format. For compressed format, such JPEG, he should use same compression ratio. That is objective. The result may be just slightly different, but it should be that. In his case, he read an JPEG image with unknown compression ratio, we write that JPEG to another JPEG with rate 1. So, two JPEGs in this case are different at all.
The PSNR does not depends on the format. You can use it on any two images, as long as the formula can be applied.
The information lost by 1 and 2 above was already lost when the first JPEG compression was done (please note that the input file is JPEG). So the initial file does not have this information in it in the first place and it cannot be lost at re-compression with a higher quality factor.
Robert Alexandru Dobre Sure, PSNR/SSIM don't depend on format. But some formats will reduce image quality so it will affect to the value of PSNR/SSIM. If compare two images in PNG and BMP or TIFF, you are right. But for JPEG or GIF, the result will be failed. Just simple example, you convert an image 32 bit bmp to GIF, JPG, the quality will down. So it is the best way you should use same format. The result will be objective.
For above example, he read JPEG with rate a, wrote to JPEG with rate b. If he wrote to PNG, I agree with you, the PSNR may be inf. But he wrote out JPEG with different ratio, the PSNR of two JPEGs are not infinity. You can check it !
Robert Alexandru Dobre
I acknowledge my mistake, it is I1=imread('new_lena.jpg');
Function imwrite(I,'nlena.png','Quality',100); doesn't work. Perhaps 'Quality' argument work with lossy jpeg compression only.
Dang N. H. Thanh
Thank you for providing a detailed explanation.
I tried writing original JPEG with Quality = 100 using JPEG format and achieved PSNR 60.6560.
Assuming original JPEG was compressed with an unknown compression rate 'c'; when I write back the JPEG with Quality=100; there is no information loss during quantization if I am understanding it right.
Only place where information loss happen is during chrominance subsampling and rounding of coefficients in DCT domain of image.
Robert Alexandru Dobre I believe that these above processing should change the information in the new image irrespective of if the original image is in JPEG or any other format.
>> I = imread('lena.jpg');
>> imwrite(I,'nlena.jpg','Quality',100);
>> J = imread('nlena.jpg');
>> psnr(I,J)
ans =
60.6560
>>
----------------------------------------------------------
>> I = imread('lena.jpg');
>> imwrite(I,'nlena.png','Quality',100);
Error using writepng>CheckTextItem (line 378)
Text chunk must be a string.
Error in writepng (line 258)
item = CheckTextItem(unmatched.(param_name));
Error in imwrite (line 472)
feval(fmt_s.write, data, map, filename, paramPairs{:});
>>
Thank you all for having this informative discussion on the topic. Best wishes to all.
- Badges
- Science topic