local tone mapping result.
My Algorithm for radiance map reconstruction is almost same as the one described in Debevec and Malik 1997 paper. The size of my A is same as the one described in lecture slide but 2pix shorter than the paper one. Atfer having solved g, I apply g on images of different exposures and take the weighted sum of those g(x) - ln(t). Before output, I use exp(hdr) to output linear scale radiance map
My global tone mapping operation is simply the log image of the linear radiance map. For local tone mapping, I calculated the Intencity manualy by multiply R/G/B channel by 0.2989/0.5870/0.1140. I implemented the bilateral filter in brute foce way but it usually take less than 1min (with space sigma =16). I am using range sigma =0.5 because the pixel radiance is in log scale.
![]() ![]() |
![]() ![]() ![]() ![]() |
![]() ![]() ![]() |
![]() ![]() ![]() |
![]() ![]() ![]() |
![]() ![]() ![]() |
![]() ![]() ![]() |
![]() ![]() ![]() |
![]() ![]() ![]() |
![]() ![]() ![]() |