Building CNNLSTM model for Sports Activity Recognition from Videos
>> YOUR LINK HERE: ___ http://youtube.com/watch?v=Po286MSk8_4
Our objective is to recognize the sports activity from the given trimmed videos. Activities is either bowling,longjump or highjump • Code Available here: • https://colab.research.google.com/dri... • Data: • https://drive.google.com/drive/folder... • • Input / Data set: Sports Video UCF 101 dataset. • CNN input:(#VideoFramesamples,#pixels,#pixels,#channels[ex. 'rgb']) • LSTM input:(#Videosamples,#timesteps,#features) • Expected Output: Find and Labelled the Sub-Activity Present in Video • CNN output: frame activity • LSTM output: Video acitvity • #trainingvideos:83 • #trainingVideoFrames:830 • #timesteps:10
#############################
