search

tomar840 / two-stream-fusion-for-action-recognition-in-videos

90 stars 28 forks source link

Features extraction at frame level #4

Closed pinkfloyd06 closed 6 years ago

pinkfloyd06 commented 6 years ago

Hello @tomar840 ,

Let me first thank you for your work.

Can l extract features representation (from last fully connected layer) for each frame (rather than for the whole video ?

Thank you a lot

tomar840 commented 6 years ago

Hi @pinkfloyd06 , This convolutional two stream fusion is only useful when you are leveraging the spatio-temporal features of the video. There is fusion of spatial and temporal features before fully connected layers, so you won't be able to extract features at the frame level.