Patrice Ferlet
Nov 4 · 1 min read

Interesting, that’s the same we are doing at Smile (where I’m working) — we are working on sign language recognition.

Are you sure that your “input” shape is well formed ? Maybe you can show me the first layer code.

It should be more or less:

model.add(keras.layers.TimeDistributed(xxx, input_shape=(5,224,224,3))

where xxx is the input layer or input model you want to distribute.

EDIT: I understand that you’re working with ImageDataGenerator that is not made for that. You’ll need to use your own generator. I will publish the next part soon, there is a simple VideoFrameGenerator that I developed. It may help you

    Patrice Ferlet

    Written by

    R&D engineer for Innovation Dept. at Smile.eu group. Coding in Golang, Python — Devops, Machine Learning, cloud computing…

    Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
    Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
    Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade