Smooth Hue Transition Interpolation in Logarithmic Exposure Space

• Again, for convenience, figure 1 is duplicated here

• Interpolation of green pixels : same as in bilinear interpolation. Also note interpolation of green pixels has to be done before interpolations of red/blue pixels.

• Key step : After the green pixel interpolations are performed,  transfer the original pixel data together with the interpolated green pixel values from linear exposure space into logarithmic exposure space. Let's denote these log values ad G_log, B_log and R_log, respectively.

• Interpolation of red/blue pixels in log space :

• In logarithmic space, define blue "hue value" as : B_log- G_log . And red "hue value" can be analogously defined.
 
• Considering the interpolation of blue pixel values : there are three different cases of blue pixel value interpolations.
 
• estimating blue pixel value at the green position and the adjacent blue pixels are on left and right : e.g. B7_log = G7_log +  (B6_log - G6_log  + B8_log - G8_log) / 2
 
• estimating blue pixel value at the green position and the adjacent blue pixels are on top and bottom : e.g. B13_log = G13_log + (B8_log - G8_log + B18_log - G18_log) / 2
 
• estimating blue pixel value at the red postion : e.g. B12_log = G12_log + (B6_log -  G6_log + B8_log - G8_log + B16_log - G16_log + B18_log - G18_log) / 4
 
• Interpolation of red pixel values can be carried out analogously

• Finally, transfer the all pixel values back into linear exposure space

• For reference, refer to [3][11][13]