topic Re: ippiResize float in Intel® Integrated Performance Primitives

ippiResize float

atlastraeger — Mon, 11 Sep 2006 23:46:17 GMT

Hello forum.

I have used ippiResize_8u_C1R to resize an image with success, however, if I try to resize a floating-point image with ippiResize_32f_C1R, 2/3 of the image is simply black (i.e. the image array consists of 0.0). For simpilicity, I converted the same image to from 8u to 32f, and then applied ippiResize_32f_C1R with the same parameters as ippiResize_8u_C1R.

What do I do wrong?

Thanks in advance.

Re: ippiResize float

Vladimir_Dudnik — Fri, 15 Sep 2006 04:58:41 GMT

Hello,

make sure you set step parameter correctly. Remember, it is number of bytes between adjacent lines, so for 8u_C1 image it can be equal to image width whereas for 32f_C1 image it is width*sizeof(Ipp32f).

Regards,
Vladimir

Re: ippiResize float

atlastraeger — Fri, 15 Sep 2006 18:54:32 GMT

Thank you very much, that was indeed the reason why it didn't work.

Re: ippiResize float

Vladimir_Dudnik — Sat, 16 Sep 2006 01:26:55 GMT

Great, hope you'll find IPP libraries useful for your tasks

Vladimir