Scaling Video Analytics on Constrained Edge Nodes

Canel, Christopher; Kim, Thomas; Zhou, Giulio; Li, Conglong; Lim, Hyeontaek; Andersen, David G.; Kaminsky, Michael; Dulloor, Subramanya R.

doi:10.48550/arxiv.1905.13536

Cited by 8 publications

(10 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Authors in [23] address the issue of network bandwidth consumption. They design FilterForward which reduces the network bandwidth consumption by transmitting only the relevant frames to the datacenter.…”

Section: Related Workmentioning

confidence: 99%

Towards a resource efficient and privacy-preserving framework for campus-wide video analytics-based applications

Gupta

Prabhat

2022

Complex Intell. Syst.

View full text Add to dashboard Cite

Video surveillance and analytics solutions based on Artificial Intelligence (AI) are increasingly being deployed across industries, including academia. There are a number of use-cases for campus-wide video analytics applications. Detecting events of interest in real-time and generating alerts is a core requirement for such applications, making them both network and compute intensive. Thus, the underlying framework needs to be resource optimized in terms of latency, compute and storage requirements for a multitude of video applications. Increasingly privacy concerns have been voiced against the pervasive deployment of video analytics-based applications. Thus, protecting the privacy of students and staff in a campus setting shall be a major design consideration for such systems going forward. This paper presents a resource optimized and privacy preserving framework for campus-wide video analytics applications. Several use-cases are presented and early results from the deployment of the proposed framework establish its feasibility and effectiveness.

show abstract

Section: Related Workmentioning

confidence: 99%

Towards a resource efficient and privacy-preserving framework for campus-wide video analytics-based applications

Gupta

Prabhat

2022

Complex Intell. Syst.

View full text Add to dashboard Cite

show abstract

“…Video. We use the night-street dataset [10] to support queries over classification tasks. Each video frame has a Boolean label indicating whether or not it contains a car.…”

Section: Datasets and Proxy Modelsmentioning

confidence: 99%

On Efficient Approximate Queries over Machine Learning Models

Ding¹,

Amer-Yahia²,

Lakshmanan³

2022

Preprint

View full text Add to dashboard Cite

The question of answering queries over ML predictions has been gaining attention in the database community. This question is challenging because the cost of finding high quality answers corresponds to invoking an oracle such as a human expert or an expensive deep neural network model on every single item in the DB and then applying the query. We develop a novel unified framework for approximate query answering by leveraging a proxy to minimize the oracle usage of finding high quality answers for both Precision-Target (PT) and Recall-Target (RT) queries. Our framework uses a judicious combination of invoking the expensive oracle on data samples and applying the cheap proxy on the objects in the DB. It relies on two assumptions. Under the Proxy Quality assumption, proxy quality can be quantified in a probabilistic manner w.r.t. the oracle. This allows us to develop two algorithms: PQA that efficiently finds high quality answers with high probability and no oracle calls, and PQE, a heuristic extension that achieves empirically good performance with a small number of oracle calls. Alternatively, under the Core Set Closure assumption, we develop two algorithms: CSC that efficiently returns high quality answers with high probability and minimal oracle usage, and CSE, which extends it to more general settings. Our extensive experiments on five real-world datasets on both query types, PT and RT, demonstrate that our algorithms outperform the state-of-the-art and achieve high result quality with provable statistical guarantees.

show abstract

“…They can be categorized into client/edge-driven methods and clouddriven methods. Representative client/edge approaches include NoScope [12], FileterForward [13], etc. They use lightweight and customized DNN models to improve the processing throughput with a lower inference time.…”

Section: B Related Workmentioning

confidence: 99%

“…In general, these systems fall into two categories, the client-driven, and the cloud-driven methods, respectively. The client-driven method runs small models [12], [13] or simple frame differencing algorithms [7] on resourcelimited client devices to filter video frames, only sending regions potentially having target objects to clouds for further processing. These methods suffer from missing important regions or sending redundant information due to the simple techniques used [11].…”

Section: Introductionmentioning

confidence: 99%

A Serverless Cloud-Fog Platform for DNN-Based Video Analytics with Incremental Learning

Zhang,

Shen,

Huang

et al. 2021

Preprint

View full text Add to dashboard Cite

Deep neural networks (DNN) based video analytics have empowered many new applications, such as automated retail and smart city. Meanwhile, the proliferation of fog computing systems provides system developers with more design options to improve performance and save cost. To the best of our knowledge, this paper presents the first serverless system that takes full advantage of the client-fog-cloud synergy to better serve the DNN-based video analytics. Specifically, the system aims to achieve two goals: 1) Provide the optimal analytics results under the constraints of lower bandwidth usage and shorter roundtrip time (RTT) by judiciously managing the computational and bandwidth resources deployed in the client, fog, and cloud environment. 2) Free developers from tedious administration and operation tasks, including DNN deployment, cloud and fog's resource management. To this end, we design and implement a holistic cloud-fog system referred to as VPaaS (Video-Platformas-a-Service) to execute inference related tasks. The proposed system adopts serverless computing to enable developers to build a video analytics pipeline by simply programming a set of functions (e.g., encoding and decoding, and model inference). These functions are then orchestrated to process video streaming through carefully designed modules. To save bandwidth and reduce RTT at the same time, VPaaS provides a new video streaming protocol that only sends low-quality video to the cloud. The state-of-the-art hardware accelerators and high-performing DNNs deployed at the cloud can identify regions of video frames that need further processing at the fog ends. At the fog ends, misidentified labels in these regions can be corrected using a light-weight DNN model. To address the data drift issues, we incorporate limited human feedback into the system to verify the results and adopt incremental machine learning to improve our system continuously. The evaluation of our system with extensive experiments on standard video datasets demonstrates that VPaaS is superior to several state-of-the-art systems: it maintains high accuracy while reducing bandwidth usage by up to 21%, RTT by up to 62.5%, and cloud monetary cost by up to 50%. We plan to release VPaaS as open-source software to facilitate the research and development of video analytics.

show abstract

Scaling Video Analytics on Constrained Edge Nodes

Cited by 8 publications

References 15 publications

Towards a resource efficient and privacy-preserving framework for campus-wide video analytics-based applications

Towards a resource efficient and privacy-preserving framework for campus-wide video analytics-based applications

On Efficient Approximate Queries over Machine Learning Models

A Serverless Cloud-Fog Platform for DNN-Based Video Analytics with Incremental Learning

Contact Info

Product

Resources

About