Yunhua468 / Audio-Visual-Emotion-and-Sentiment-Research

Deep Neural Network and its application with TensorFlow project
2 stars 0 forks source link

How to merge video and audio features? #4

Open Yunhua468 opened 4 years ago

Yunhua468 commented 4 years ago

how to merge video and audio features, there are several ways as my understanding:

  1. extract features from the original files, then merge them directly, then input to a deep nueral network
  2. Or after extracting features from the original files, use deep nueral network to extract deep features, then combine them to input a deep neural network/machine learning alg to do classification
  3. Or as what Enis recommanded, classify them separately then choose a higher accuracy result For what we learned on class, I prefer the 2), or we could do it separately to see the result. What do you guys think about?
EnisBerk commented 4 years ago

I think we should try all of these and compare the performances