In the paper it is mentioned that the average value of the ACD for ground truth videos of the same person is 0.74·10^−4. It is also mentioned that the metric is obtained by euclidean distance between the still image representation, obtained using OpenFace, from the representation of the generated frames.
However the feature distance computed by me using Openface on a single person's ground truth video does not include values in the range 10^-4. Could you please share the exact formula or the code used for obtaining the ACD values reported in the paper.
In the paper it is mentioned that the average value of the ACD for ground truth videos of the same person is 0.74·10^−4. It is also mentioned that the metric is obtained by euclidean distance between the still image representation, obtained using OpenFace, from the representation of the generated frames. However the feature distance computed by me using Openface on a single person's ground truth video does not include values in the range 10^-4. Could you please share the exact formula or the code used for obtaining the ACD values reported in the paper.