https://community.intel.com/t5/Intel-Integrated-Performance/Doubt-about-Image-operation/m-p/947265#M18280 <DIV>Hello,</DIV> <DIV></DIV> <DIV>I have a question about image operation. I want to do the Radon Transform. That Radon transform consists to make the sum of the columns pixel values, and put the value into another image, and then rotate. IPP have rotate funtion, and this it's no a problem.</DIV> <DIV></DIV> <DIV>If I use an image depth 8U, I lose information because the sum it's higher than 255.</DIV> <DIV></DIV> <DIV>But if I use an image depth 32F, I don't know how to sum the pixel values because each pixel have 4 bytes.</DIV> <DIV></DIV> <DIV>I would be very grateful if anyone could help me.</DIV> <DIV></DIV> <DIV></DIV> <DIV>Thank you.</DIV> <DIV></DIV> <DIV></DIV> <DIV></DIV> <DIV>Enric </DIV> Wed, 08 Feb 2006 03:35:54 GMT enricbosch2002 2006-02-08T03:35:54Z https://community.intel.com/t5/Intel-Integrated-Performance/Doubt-about-Image-operation/m-p/947265#M18280 <DIV>Hello,</DIV> <DIV></DIV> <DIV>I have a question about image operation. I want to do the Radon Transform. That Radon transform consists to make the sum of the columns pixel values, and put the value into another image, and then rotate. IPP have rotate funtion, and this it's no a problem.</DIV> <DIV></DIV> <DIV>If I use an image depth 8U, I lose information because the sum it's higher than 255.</DIV> <DIV></DIV> <DIV>But if I use an image depth 32F, I don't know how to sum the pixel values because each pixel have 4 bytes.</DIV> <DIV></DIV> <DIV>I would be very grateful if anyone could help me.</DIV> <DIV></DIV> <DIV></DIV> <DIV>Thank you.</DIV> <DIV></DIV> <DIV></DIV> <DIV></DIV> <DIV>Enric </DIV> Wed, 08 Feb 2006 03:35:54 GMT https://community.intel.com/t5/Intel-Integrated-Performance/Doubt-about-Image-operation/m-p/947265#M18280 enricbosch2002 2006-02-08T03:35:54Z https://community.intel.com/t5/Intel-Integrated-Performance/Doubt-about-Image-operation/m-p/947266#M18281 <DIV>Hi, Enric,</DIV> <DIV></DIV> <DIV>IPP nas no functions for summation of rows or columns of 8u images</DIV> <DIV>But there are such functions for 16s data (Vol 1 pp 8-9 - 8-12)</DIV> <DIV><FONT face="Courier" size="2"> <P>ippsSumColumn_16s32s_D2Sfs</P><FONT face="Courier" size="2"> <P>ippsSumColumn_16s32f_D2</P></FONT></FONT></DIV> <DIV> <DIV><FONT face="Courier" size="2"> <P>ippsSumRow_16s32s_D2Sfs</P><FONT face="Courier" size="2"> <P>ippsSumRow_16s32f_D2</P></FONT></FONT></DIV>Note: these functions require row step in elements not in bytes</DIV> <DIV></DIV> <DIV>Alexander</DIV> Thu, 09 Feb 2006 17:46:55 GMT https://community.intel.com/t5/Intel-Integrated-Performance/Doubt-about-Image-operation/m-p/947266#M18281 Intel_C_Intel 2006-02-09T17:46:55Z