Name
Abstract | ||
|---|---|---|
Media analytics (MA) usages are growing rapidly in various application domains. A common pipeline consists of the elements of communication, media processing and content analysis. However, designing systems with both high throughput and scalability is a difficult problem. In this paper, we present a solution which is based on Intel® architecture and has significant innovations on both the hardware and software sides. In contrast with existing accelerators, the new acceleration card supports the entire MA pipeline on the card and has loose coupling with the host. Experiments have shown superior abilities of the system in density, scalability and cost efficiency when used for different video analytics tasks. Deployment of the card in smart city and live streaming use cases is also described. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1109/MIPR49039.2020.00057 | 2020 IEEE Conference on Multimedia Information Processing and Retrieval (MIPR) |
Keywords | DocType | ISBN |
Media Analytics,Deep Learning,Inference,AI Accelerator,Edge Computing | Conference | 978-1-7281-4273-9 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jingyi Jin | 1 | 0 | 0.68 |
| Tong Zhang | 2 | 0 | 0.68 |
| Kevin Cone | 3 | 0 | 0.34 |
| Beryl Xu | 4 | 0 | 0.34 |