Confusion whether the calculated compression ratio is correct or not.

I am applying compression on an image of 40KB(350x500x3 uint8) to get new image of 18KB(348x500x3 uint8) .I have used the following code for finding the compression ratio-

z='xyz.jpg';
      z1='out.jpg';
      k=imfinfo(z);
      k1=imfinfo(z1);
      ib=k.Width*k.Height*k.BitDepth/8;
      cb=k1.Width*k1.Height*k1.BitDepth/8;
      cr=ib/cb;

The value of cr comes to be 1.005747126436782 .

I don't know whether the value of cr is correct or not.Can anyone verify it? Also i would like to know how it is manually calculated via the formula.

[Guillaume]

Depends on your definition of compression ratio.

I assume the byte sizes (40 kB vs 18 kB) you give are the file size (which will depend on storage format) whereas your formula compares the memory size (~525 kB vs ~522 kB).

[Nidhi Kumari]

So what is the correct coding approach according to the byte size?

[Walter Roberson]

In https://www.mathworks.com/matlabcentral/answers/433311-how-to-compute-number-of-dedicated-bits-in-arithmetic-coding-in-matlab#answer_350087 I mention some of the perils of relying on file size to calculate compression ratio.