topic thanks for reporting this in Intel® Integrated Performance Primitives

Misleading histogram example

Gramkow__Claus — Tue, 18 Sep 2018 07:24:51 GMT

The example given for the use of histograms (https://software.intel.com/en-us/ipp-dev-reference-histogram) seems misleading. The upper limit is set to 255 and the criterion for an intensity to belong in a bin is stated to be pLevels<=pSrc(x, y)<pLevels(k+1). Therefore, if the randomized image contains the value 255, the sum of frequencies will not reach 64 = 8x8, which very much seems to be the intention of the example.

See the thread "ippiHistogram

Adriaan_van_Os — Wed, 19 Sep 2018 05:26:49 GMT

See the thread "ippiHistogram_8u_C1R problem" https://software.intel.com/en-us/forums/intel-integrated-performance-primitives/topic/676143 and my comment on it. There never was a reply.

Regards,

Adriaan van Os

thanks for reporting this

Chao_Y_Intel — Fri, 21 Sep 2018 03:29:49 GMT

thanks for reporting this issue. I tracked in our bug database, so we can fix it in the future release. regards, Chao