How to program a RNN model with image sequences of varying length in keras for regression? (Just assume the output is some predefined continuous values)

I've read up things about training RNN on sequences with varying lengths like: here and here. But they all refers to numerical sequences. How can I deal with it in terms of image sequence of varying length in keras? Let's say I have sequences of varying length where each image sized 9*16 pixels. How to implement this in Keras?

Note: I also need the model to output intermediate values in hidden state.

Thanks in advance.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.