Evaluating text-camera coherence

robincourant / DIRECTOR

MIT License

54 stars 4 forks source link

Evaluating text-camera coherence #5

Closed AdityaDhaulakhandi closed 1 week ago

AdityaDhaulakhandi commented 1 week ago

Hi, can you point me to the code where motion tagging is performed for a generated trajectory to derive Classifier Precision (C-P), Classifier Recall (C-R) and Classifier F1-Score (C-F1)? Thanks!

robincourant commented 1 week ago

Hi, You'll find the metric code in: https://github.com/robincourant/DIRECTOR/blob/main/src/metrics/modules/caption.py