Recently, additional child safety concerns have been raised after alleged incidents of abuse and neglect were reported at a children’s foster home, and called for better monitoring and enforcement of the law. Meanwhile, the Omicron outbreak has also put a heavy strain on public health services and isolation facilities, which have seen a surge in cases of infection.
Therefore, video surveillance appears to be an alternative that will help alleviate the shortage of medical staff in both clinical and older facilities during the pandemic situation. Still, security and privacy remain some of the video surveillance concerns that need to be addressed. With this in mind, the Data Science Lab at the Department of Statistics and Actuarial Science (SAAS) at the University of Hong Kong (HKU) has developed several real-time video analytics apps that use the first edge computing AI chip donated in Hong Kong by one Hong Kong-based technology companies, enabling them to recognize human body movements (e.g. walking, falling, leaving the room, etc.) and facial expressions (e.g. crying, screaming, etc.).
Using bounding box detection, object detection, and motion classification techniques, the team built the apps using ResNet-32, a deep learning neural network specified for image recognition, and Flasks, the Python programming language framework for video analysis and used an average of over 5,000 images collected from the internet to train each model.
By densifying the data center network, the edge computing AI chip enables faster data processing, better data security and efficient control over continuous operations – it is 10 times faster than the market’s edge-based AI chip in terms of computing power.
Because the video analytics is done in the AI chip itself, the apps don’t need to run on a cloud computing platform, overcoming the security and privacy concerns of streaming video over the internet. Moreover, the AI chip can be implemented in any device like robot pet, surveillance camera, etc. It can be used in home care/child care/elderly care centers, offices, shopping malls or hotels for risk detection and personal care monitoring.
In the case of child abuse in children’s homes, the video analysis apps can be used to monitor childcare. Also, the apps can assist nurses in monitoring patient risks in isolation wards, alert them when patients leave the ward, ask for help or intentionally self-extubate their endotracheal tubes, and help retirement homes to identify risks in elderly care such as falling, crying, etc .
The CEO of the tech company that donated the chip is the first edge computing AI chip developed by the Hong Kong firm. He thanked the Data Science Lab for deploying the reinvented AI chip with its novel real-time video analytics apps that overcome the security and privacy concerns of video analytics and surveillance.
In addition to security monitoring, the AI chip can also be used in the processing and analysis of big data in finance and medicine, investment decision-making and disease diagnosis, virtual reality, robotic automation, smart home, physical and cognitive training, medicine and gene discover.
The deputy director of the HKU SAAS Data Science Lab explained that the lab is currently researching a novel parallel and distributed AI algorithm and plans to deploy it in an environment of clustered edge computing AI chips. In this way, big data analysis can be performed in real time with supercomputing performance, and is a breakthrough in AI development.
About the HKU SAAS Data Science Lab
The project team of the HKU Data Science Lab consists of the head of the Department of Statistics and Actuarial Science (SAAS) Professor Guosheng YIN, the director of the SAAS Data Science Lab Dr. Eddy LAM, the deputy director Dr students of the department.