DEEP LEARNING PROJECT BASED ON HUMAN ACTIVITY RECOGNITION
Sarumathi G1, Susmitha T2, Shivani S3, Mr. Sathish S4
1Department of Computer Science and Business Systems & Bannari Amman Institute of Technology
2 Department of Electronics and Communication Engineering & Bannari Amman Institute of Technology
3 Department of Computer Science and Engineering & Bannari Amman Institute of Technology
4 Department of Computer Science and Business Systems & Bannari Amman Institute of Technology
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The main objective of this paper is to implement human activities by performing some of the specified tasks. In order to classify a time series, the challenge of human activity recognition requires data from a number of timesteps. The methodology that would be followed in the paper is using deep learning and classification methods to make predictions with the model. To extract the local features of each frame, a convolutional neural network and Long Short Term Memory (LSTM) of the deep learning algorithm will be used. This implementation will help to monitor human activity and ensure their safety by sending a message notification through their mobile phones using Twilio packages that is a library. This is a programmable messaging software application that will send notifications from the prevention of danger and the monitoring process will be easy to know and convenient. The project will be developed as an application for the users to interact as well as analyze real-time data analysis.In this paper,we will discuss about how human activity recognition can effectively be employees to ensure safety and security of elderly people, women and also in burglary prevention.The complete implementation proposal along with the technical specifications have been designed so as to implement best possible solutions.
Keyword - CNN, LSTM, Deep Learning, Twilio packages, User Interface.