Human Activity Recognition Using LRCN

Using the capabilities LRCN it obtained to increase the multimedia, recognize human movement in videos. HAR is a one ongoing research field. On various benchmark datasets, new records have been established using video as a series of frames. Convolutional neural networks (CNN) and long short-term memories were proposed in this (LSTM). CNN network is applied to get feature vectors for each videos and LSTM network is use to classify the videos. Standard activities contains many activities horse racing, taichi, walking with dog, basketball, swing, etc. In this proposed system we have achieved 99.5% of Accuracy.