Human Action Detection using Image Data and Deep Learning Algorithms

Abstract

Human action detection has emerged as one of the critical research areas both in computer vision and deep learning, inspired by various applications related but not limited to surveillance and human-computer interaction. This paper, therefore, sets out to perform a review study on how deep learning algorithms have been applied in detecting and recognizing actions from image sequences. It starts with a review of the literature on state-of-the-art action recognition techniques and then discusses their relative advantages and limitations. The proposed system makes use of convolutional neural networks combined with long short-term memory for a superior understanding of spatial and temporal features in images. In the experimental evaluation of the model, a dataset of human actions was used, and the results showed that the proposed approach provides greater accuracy compared to traditional approaches. This paper provides a discussion on the possible improvements and applications of this method in real-world scenarios.

DOI - https://doi.org/10.65525/SVUP.9788199778009.2026.152-162