Human Activity Recognition Using EfficientNet-B2 Deep Learning Model

Human activity recognition refers to the automated process of categorizing and classifying human activities that is derived from sensor data or visual input. They are widely used in research, benchmarking various algorithms, and building real-world applications that benefit from understanding and predicting human activity, such as healthcare, fitness tracking, assistive technology, and intelligent environments. EfficientNetB2 is well known for its superiority in performance and efficiency, making it an ideal option for tasks requiring HAR. This study demonstrates the efficacy of the EfficientNetB2 model in precisely classifying a range of human activities using data gathered from several sensors. Our goal is to evaluate the model's performance in classifying different human activities. The results clearly show that this model gave the highest accuracy of 75 percent for 9 epochs