How do I convert a video to the data format required for this model?

[She-yh]

I've got the bimodal_weights_emotion.hdf5 model from baseline, and the model can recognize the emotion from MELD dataset well. But I don't know how to recognize emotion from my own video. How do I convert a video to the data format required for this model?

I'm a beginner in multimodal emotion recognition. I'd really appreciate it if you could give me some tips