2.Objective
Artificial intelligence (AI) and computer learning have improved rapidly in the recent decade. However, the capability of applying this technology is limited by the capacity of computing resources and infrastructures. It requires accelerated computing resources, such as graphics processing units (GPUs), in order to perform expensive processing tasks in real time. Moreover, for image processing and computer vision tasks, the system requires a high-bandwidth network to transfer rich image data from acquiring units, i.e. cameras, to processing units which is usually a centralized server. To solve such the issue, the most recent AI technologies tend to focus on edge processing and decentralized cloud computing.
In order to accelerate the adoption of AI technologies for public safety and security in Thailand, we propose a design and implementation of a distributed computing platform that can provide high performance video analytics with lower cost for deployment. We implement an AI camera, called AI in-a-Box (AIB), which combines network camera with the capabilities of AI and embedded technologies. Moreover, we also propose a design of cloud computing platform which can provide high-performance AI computing service and interactive user interface.
3.Sponsor
Thailand Science Research and Innovation
4.Budget
15 million baht
5.Photos/Videos
