Hey,
I'm trying to classify a video action sequence which contains 5 sub-actions in the whole sequence. I thought maybe I can use one-hot encoding as label like [0 1 1 0 0] to see which sub-actions were not performed. Can I use these features as input to the network for testing video sequence and output label would be something like [1 0 0 1 0] for example.
Thanks
Hey,
I'm trying to classify a video action sequence which contains 5 sub-actions in the whole sequence. I thought maybe I can use one-hot encoding as label like [0 1 1 0 0] to see which sub-actions were not performed. Can I use these features as input to the network for testing video sequence and output label would be something like [1 0 0 1 0] for example.
Thanks