Researchers at the University of Virginia’s School of Engineering and Applied Science have developed a new AI-driven video analyzer called the Semantic and Motion-Aware Spatiotemporal Transformer Network (SMAST). This cutting-edge system is capable of recognizing and predicting complex human actions, making it a valuable tool for various industries.
SMAST’s ability to process intricate video footage lies in its integration of a multi-feature selective attention model and a motion-aware 2D positional encoding algorithm. This enables the AI to focus on essential elements within a scene, such as a person in motion, while ignoring extraneous details. The system can distinguish between various actions, making it effective in recognizing complex behaviors.
The potential applications of SMAST are vast, with researchers suggesting its use in surveillance, healthcare, and autonomous driving. In the security and surveillance sectors, SMAST could enhance public safety by detecting threats in real-time and identifying suspicious behavior. In healthcare, the technology could track patients’ movements for better motion analysis during rehabilitation or surgery.
The researchers at the University of Virginia believe that SMAST’s ability to handle chaotic, unedited footage sets it apart from other AI systems. The tool has been tested on various academic benchmarks and has performed well, demonstrating its potential for real-time action detection in challenging environments.
According to Professor Scott T. Acton, the chair of the Department of Electrical and Computer Engineering at the University of Virginia, SMAST has the potential to prevent accidents, improve diagnostics, and even save lives. This groundbreaking AI technology is a significant advancement in the field of visual data analysis and has promising implications for the future of various industries.
Source
Photo credit www.techradar.com