ai4r / Gesture-Generation-from-Trimodal-Context

Speech Gesture Generation from the Trimodal Context of Text, Audio, and Speaker Identity (SIGGRAPH Asia 2020)
Other
245 stars 35 forks source link

How to use FGD for custom input? #50

Closed Rohith-coder1 closed 1 year ago

youngwoo-yoon commented 1 year ago

Hello, You need to train a feature extractor (or you can use raw data if you want), then collect features from original data and generated data, and then you can calculate the frechet distance between two sets of features. This overall process is in https://github.com/ai4r/Gesture-Generation-from-Trimodal-Context/blob/master/scripts/model/embedding_space_evaluator.py

I also used FGD for other gesture data. Please check this out too. https://github.com/Svito-zar/genea_numerical_evaluations/tree/master/FGD