TY - GEN
T1 - AdaMask
T2 - 30th ACM International Conference on Multimedia, MM 2022
AU - Liu, Shengzhong
AU - Wang, Tianshi
AU - Li, Jinyang
AU - Sun, Dachun
AU - Srivastava, Mani
AU - Abdelzaher, Tarek
N1 - Acknowledgments Research reported in this paper was sponsored in part by the Army Research Laboratory under Cooperative Agreement W911NF-17-20196, NSF CNS 20-38817, and the Boeing Company.
PY - 2022/10/10
Y1 - 2022/10/10
N2 - This paper presents AdaMask, a machine-centric video streaming framework for remote deep neural network (DNN) inference. The objective is to optimize the accuracy of downstream DNNs, offloaded to a remote machine, by adaptively changing video compression control knobs at runtime. Our main contributions are twofold. First, we propose frame masking as an effective mechanism to reduce the bandwidth consumption of video stream, which only preserves regions that potentially contain objects of interest. Second, we design a new adaptation algorithm that achieves the Pareto-optimal tradeoff between accuracy and bandwidth by controlling the masked portions of frames together with conventional H.264 control knobs (eg. resolution). Through extensive evaluations on three sensing scenarios (dash camera, traffic surveillance, and drone), frame masking saves the bandwidth by up to 65% with < 1% accuracy degradation, and AdaMask improves the accuracy by up to 14% over the baselines against the network dynamics.
AB - This paper presents AdaMask, a machine-centric video streaming framework for remote deep neural network (DNN) inference. The objective is to optimize the accuracy of downstream DNNs, offloaded to a remote machine, by adaptively changing video compression control knobs at runtime. Our main contributions are twofold. First, we propose frame masking as an effective mechanism to reduce the bandwidth consumption of video stream, which only preserves regions that potentially contain objects of interest. Second, we design a new adaptation algorithm that achieves the Pareto-optimal tradeoff between accuracy and bandwidth by controlling the masked portions of frames together with conventional H.264 control knobs (eg. resolution). Through extensive evaluations on three sensing scenarios (dash camera, traffic surveillance, and drone), frame masking saves the bandwidth by up to 65% with < 1% accuracy degradation, and AdaMask improves the accuracy by up to 14% over the baselines against the network dynamics.
KW - deep neural networks (DNN)
KW - edge offloading
KW - video streaming
UR - http://www.scopus.com/inward/record.url?scp=85151129155&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85151129155&partnerID=8YFLogxK
U2 - 10.1145/3503161.3548033
DO - 10.1145/3503161.3548033
M3 - Conference contribution
AN - SCOPUS:85151129155
T3 - MM 2022 - Proceedings of the 30th ACM International Conference on Multimedia
SP - 3035
EP - 3044
BT - MM 2022 - Proceedings of the 30th ACM International Conference on Multimedia
PB - Association for Computing Machinery
Y2 - 10 October 2022 through 14 October 2022
ER -