In the paper MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation, I find the g_n = M @ g_r, and the conversion matrix M = S @ R。
but in the denormalize process, it seems to use the R not the M to denormalize.
Did I understand something wrong?
In the paper MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation, I find the g_n = M @ g_r, and the conversion matrix M = S @ R。 but in the denormalize process, it seems to use the R not the M to denormalize. Did I understand something wrong?
