Open Crescent-Saturn opened 2 years ago
@Crescent-Saturn , could you please contribute?
@Crescent-Saturn , could you please contribute?
Hi, @nmanovic , thanks for consideration. Alright, I would like to have a try. Once completed, I will make a PR.
My actions before raising this issue
In some cases, raw images are hard to annotated due to their contrasts, brightness, etc. It is great to see that CVAT has already integrated the Histogram equalization with OpenCV tools in the GUI. However, the default
cv.equalizeHist
function may work well as it considers the global contrast of the image. This can be improved by using the CLAHE (Contrast Limited Adaptive Histogram Equalization) algo which exists already in OpenCV.Expected Behaviour
Using CLAHE to perform histogram equalization.
Current Behaviour
Using
cv.equalizeHist
function to permform histogram equalization. #3447Possible Solution
Replace
cv.equalizeHist
bycv.CLAHE(clipLimit, tileGridSize)
andcv.CLAHE().apply()
.Where
clipLimit
andtitleGridSize
may be adjusted by adding a Trackbar.Steps to Reproduce (for bugs)
1. 1. 1. 1.
Context
Thanks for consideration!
Your Environment
git log -1
):docker version
(e.g. Docker 17.0.05):Logs from `cvat` container