Abstract | ||
---|---|---|
In recent years, convolutional networks have demonstrated unprecedented performance in the image restoration task of super-resolution (SR). SR entails the upscaling of a single low-resolution image in order to meet application-specific image quality demands and plays a key role in mobile devices. To comply with privacy regulations and reduce the overhead of cloud computing, executing SR models locally on-device constitutes a key alternative approach. Nevertheless, the excessive compute and memory requirements of SR workloads pose a challenge in mapping SR networks on resource-constrained mobile platforms. This work presents MobiSR, a novel framework for performing efficient super-resolution on-device. Given a target mobile platform, the proposed framework considers popular model compression techniques and traverses the design space to reach the highest performing trade-off between image quality and processing speed. At run time, a novel scheduler dispatches incoming image patches to the appropriate model-engine pair based on the patch's estimated upscaling difficulty in order to meet the required image quality with minimum processing latency. Quantitative evaluation shows that the proposed framework yields on-device SR designs that achieve an average speedup of 2.13x over highly-optimized parallel difficulty-unaware mappings and 4.79x over highly-optimized single compute engine implementations.
|
Year | DOI | Venue |
---|---|---|
2019 | 10.1145/3300061.3345455 | MobiCom '19: The 25th Annual International Conference on Mobile Computing and Networking
Los Cabos
Mexico
October, 2019 |
Keywords | DocType | ISBN |
Super-resolution,deep neural networks,mobile computing,heterogeneous computing,scheduling | Conference | 978-1-4503-6169-9 |
Citations | PageRank | References |
8 | 0.48 | 13 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Royson Lee | 1 | 8 | 2.85 |
Stylianos I. Venieris | 2 | 106 | 12.98 |
Łukasz Dudziak | 3 | 17 | 4.37 |
Sourav Bhattacharya | 4 | 624 | 52.45 |
Nicholas D. Lane | 5 | 4247 | 248.15 |