Real Time Human Activity Recognition with Video Classification

EasyChair Preprint 7377

Abstract

Recognition of the human activity is a very broad area of study that will aim to identify the specific movement or the action of the person. Human activity recognition is a very vast area of research and exploration This is a type of time series classification problem where data from a series of time steps is needed to properly classify the current activity. Activities are actions such as walking, eating, sleeping, reading newspapers, talking, jumping, standing, drinking and sitting. Recognizing human activity, or HAR, is a very difficult task of classifying the data points. To put this in simple words, the action of classification or prediction the activity or the action will be performed by someone is known as the activity recognition. The issue which arises here is, in the action Recognition, you will actually need the series of the knowledge-based points for the prediction of the action which is being performed accurately. So, the action-based Recognition would be in a form of the statistic classification with the draw back where it is likely that the data from the series of the timesteps to properly classify the action which is being performed properly. It also involves the prediction of the movement of an individual data and will involve the deep domain expertise and different methods from the signal method to engineer choices properly from the information therefore on the pursuit a machine learning model. Recently, the deep learning models will appreciate the convolutional neural networks and the continual neural networks which have shown the capable and even showed the successful progressive results by automatic learning options from the raw sensing element data.

Keyphrases: Artificial Intelligence, Convolutional Neural Network, Human Activity Recognition, Real videos, deep learning

@booklet{EasyChair:7377,
  author    = {S Janhavi and Chandra Sekhar Malepati},
  title     = {Real Time Human Activity Recognition with Video Classification},
  howpublished = {EasyChair Preprint 7377},
  year      = {EasyChair, 2022}}