Thank you for sharing this repo!
I am trying to use this to do Facial Recognition by comparing similarity scores. For each frame, do I use
|
indv_similarities = calc_cossim(probe_features, gallery_feature) |
or
|
fused_similarity = calc_cossim(np.expand_dims(probe_fused_feature, 0), gallery_feature) |
between a detected face and reference images?
Thank you for sharing this repo!
I am trying to use this to do Facial Recognition by comparing similarity scores. For each frame, do I use
caface/demo/visualization.py
Line 20 in e94ad02
or
caface/demo/visualization.py
Line 21 in e94ad02
between a detected face and reference images?