How to use the pre-traiend model to input the action as text in the videos

hx173149 / C3D-tensorflow

C3D is a modified version of BVLC tensorflow to support 3D ConvNets.

MIT License

588 stars 265 forks source link

How to use the pre-traiend model to input the action as text in the videos #96

Open momo1986 opened 5 years ago

momo1986 commented 5 years ago

Hello, Thanks a lot for the pre-trained model.

If I load c3d_ucf101_finetune_whole_iter_20000_TF.model in my own environment, how can I test the change of the images and output the actions from prediction_result based on the pre-trained model?

Is there any demo available?

Thanks & Regards!

highway007 commented 5 years ago

@momo1986 Here it is https://github.com/FingerRec/real_time_video_action_recognition.git