[FEA] Improve performance of existing `equalize_adapthist` function (CLAHE)

[grlee77]

Is your feature request related to a problem? Please describe.
The current implementation of equalize_adapthist is much slower than the CUDA implementation of CLAHE for uint8_t or uint16_t 2D images available in OpenCV-contrib.

Describe the solution you'd like
The current implementation matches the scikit-image implementation closely, but can likely be refactored to improve efficiency while maintaining the current behavior.