IPLUSO researchers developed an AI-based intelligent surveillance system capable of detecting suspicious activities with 99.12% accuracy
Researchers from IPLUSO developed an advanced surveillance model based on artificial intelligence, named GM_CNN3D, designed to detect suspicious human activities in complex and crowded environments. The model combines traditional motion descriptors with deep learning capabilities, improving accuracy and reducing false alarms in real-world surveillance scenarios.
The system was tested on several public datasets, including videos of crowd violence and confrontation detection, achieving accuracy rates of up to 99.12%, with robust performance under different conditions — including varying lighting levels, crowd densities and camera angles.
The proposed approach demonstrates greater reliability when compared to existing deep learning models and shows strong potential for real-time security applications and automated alert systems.
Read the Article Here
