pub fn apply_gradient_weights_f32(
out_x: &mut [f32],
out_y: &mut [f32],
out_z: &mut [f32],
mx: &[f32],
my: &[f32],
mz: &[f32],
p: &[f32],
gx: &[f32],
gy: &[f32],
gz: &[f32],
)Expand description
Apply per-direction gradient weights: out[i] = mx[i] * p[i] * mx[i] * gx[i] This is the core operation in MEDI’s regularization term