Bi-LSTM Model to Recognize Human Activities in UAV Videos using Inflated I3D-ConvNet

Human activity recognition in aerial videos is an emerging research area. In this paper, Inflated I3D-ConvNet (Inflated I3D) and Bidirectional Long Short-Term Memory (Bi-LSTM) based human action model UAV have been proposed. The initial module was pre-trained using the Kinetics-400 video dataset, which consisted of 400 classes activities around clips for each class culled from real-world arduous YouTube videos. proposed inflated built on 2D-ConvNet inflation learns extracts spatio-temporal features while leveraging architectural design Inception-V1. employs Bi-LSTM architecture classification Drone-Action dataset a smaller benchmark UAV-captured dataset. This considerably improves state-of-the-art results SoftMax classifier retains accuracy about 98.4%.