Human Activity Recognition Using Wearable Devices

“Human Activity Recognition Using Wearable Devices” is our research, which provides an extensive view on Deep Learning(DL) on mobile platforms with a focus on its embedding in wearable technology for activity recognition. This investigation investigates the distinct advantages that come along when utilizing DL on mobile devices such as improved private data handling, reduced communication overheads and lowered system costs are opposed to cloud-based versions. We explore vision-based human activity recognition(HAR), especially in healthcare applications, which are currently evolving at a breakneck pace so that we can make an exhaustive study of deep learning mechanism that can be employed for this purpose. Classified into Convolutional using techniques based on Convolutional Neural Networks(CNN), Recurrent Neural Networks(RNN) and LSTM, we investigate a range of subcategories and assess their effectiveness on a variety of benchmark datasets. Furthermore, we compare deep learning fusion techniques with handmade feature-based approaches to explore the evolution of HAR methodologies.