Researchers at the University of Virginia’s School of Engineering and Applied Science have developed a new AI system called SMAST that is capable of recognizing complex human actions in video footage. The system, which utilizes a selective attention model and a motion-aware algorithm, can effectively detect and interpret behaviors in real-time.
SMAST has potential applications in various industries, including surveillance, healthcare, and autonomous driving. In the security sector, the system can enhance public safety by identifying potential threats and suspicious behavior. In healthcare, it could be used for motion analysis during rehabilitation or monitoring patients during surgery.
The researchers at UVA have tested SMAST on several academic benchmarks and the system has shown promising results. Professor Scott T. Acton, chair of the Department of Electrical and Computer Engineering, believes that this technology can make a significant impact by preventing accidents, improving diagnostics, and potentially saving lives.
SMAST’s ability to handle chaotic and unedited footage sets it apart from other AI systems, as it can adapt to different environments and improve its action detection capabilities over time. This innovation represents a significant advancement in AI-driven visual data analysis and has the potential to revolutionize how we interact with technology in the future.
Source
Photo credit www.techradar.com
