• Vaishnavi S
• Ms. Divyashree S
• Shrusti S Rao
• Sanjana Shenoy

This project centers on the development of a human activity recognition (HAR) system utilizing Convolutional Neural Networks (CNNs) for precise identification of diverse human activities in video footage. The system architecture encompasses three key components: data pre-processing, model selection, and model evaluation. Data pre-processing involves meticulous cleaning and exploratory analysis to optimize model performance. Model selection involves the strategic choice between CNN and recurrent neural network (RNN) models. Model evaluation entails training the selected model on a video dataset and evaluating its performance on an independent test dataset. Harnessing recent advancements in HAR facilitated by Machine Learning (ML) techniques, this project thoroughly explores the potential of CNNs. By transitioning from conventional machine learning approaches to CNNs with automatic feature extraction, the framework aims to improve real-time forecasting of human behaviors and detection of anomalous activities. Targeting domains like senior care and surveillance systems, this implementation contributes to enhancing public safety measures.

Human activity recognition (HAR), Convolutional Neural Networks (CNNs), Data pre-processing, Model selection, Model evaluation Machine Learning (ML) techniques, Anomalous activity detection, Surveillance system

"Implementation Paper: Human Activity Recognition using CNN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 5, page no.319-330, May-2024, Available :http://www.jetir.org/papers/JETIRGG06051.pdf

"Implementation Paper: Human Activity Recognition using CNN", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 5, page no. pp319-330, May-2024, Available at : http://www.jetir.org/papers/JETIRGG06051.pdf