Edge-Cloud Cooperative Image Processing by Partially Streaming ROI Data for Metaverse Applications

Recently, interest in the metaverse has been increasing. In the multi-edge-based structure of the metaverse, the edge consumes a lot of power during image processing and data transmission. These problems make the metaverse system difficult to use in streaming, which leads to an increase in the edge's size and performance. We propose a collaborative edge-cloud of image processing and data transmission to reduce the edge's size and power consumption. We use a Region of Interest (ROI) and distribute edge-image processing to make streamable and runtime executable AI software. With this proposed structure, we confirmed the reduction in edge time, power consumption, and network communication.