Attention-driven BI-LSTM for Robust Human Activity Recognition and Classification

Accurate and robust Human Activity Recognition is essential for applications in surveillance, healthcare, and smart environments. However, the unpredictability and complexity of human motions provide significant challenges in obtaining the desired levels of accuracy and robustness. Conventional machine learning models, such as Decision Tree, Gaussian NB, and KNeighbors, have shown limited efficacy, with estimates of accuracy ranging from 78.3% to 89.3%. Cutting-edge techniques like as Random Forest, RBF SVC, and XGB Classifier achieve a maximum accuracy of 93.8%. We present an Attention-Driven BI-LSTM model that uses bi-directional long short-term memory networks improved with a devotion mechanism to prioritize the most important characteristics in order to overcome these constraints. The present model demonstrates exceptional performance, attaining an accuracy of 99.83%, a precision of 99.46%, a recall of 99.75%, and an F1 score of 99.85%, thereby surpassing other approaches by a substantial margin. The obtained findings validate the model's resilience and effectiveness in precisely recognizing and categorizing human actions in different fields and situations.