bilinearInterpolatePowerOf2()

u8 fl::bilinearInterpolatePowerOf2	(	u8	v00,
		u8	v10,
		u8	v01,
		u8	v11,
		u8	dx,
		u8	dy )

Definition at line 222 of file upscale.cpp.

                                                          {
    u16 dx_inv = 256 - dx; // 0 to 256
    u16 dy_inv = 256 - dy; // 0 to 256
 
    // Scale down weights to fit into u16
    u16 w00 = (dx_inv * dy_inv) >> 8; // Max value 256
    u16 w10 = (dx * dy_inv) >> 8;
    u16 w01 = (dx_inv * dy) >> 8;
    u16 w11 = (dx * dy) >> 8;
 
    // Sum of weights should be approximately 256
    u16 weight_sum = w00 + w10 + w01 + w11;
 
    // Compute the weighted sum of pixel values
    u16 sum = v00 * w00 + v10 * w10 + v01 * w01 + v11 * w11;
 
    // Normalize the result
    u8 result = (sum + (weight_sum >> 1)) / weight_sum;
 
    return result;
}

Referenced by upscalePowerOf2(), and upscaleRectangularPowerOf2().

Here is the caller graph for this function: