Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adaptive Offloading and Enhancement for Low-Light Video Analytics on Mobile Devices

Yuanyi He, Peng Yang, Tian Qin, Jiawei Hou, Ning Zhang

TL;DR

A novel enhancement quality assessment method is proposed on CAMs to evaluate the effectiveness of different enhancement algorithms for low-light videos, and a genetic-based scheduling algorithm is proposed, which can find a near-optimal solution in a reasonable time to meet the latency requirement.

Abstract

In this paper, we explore adaptive offloading and enhancement strategies for video analytics tasks on computing-constrained mobile devices in low-light conditions. We observe that the accuracy of low-light video analytics varies from different enhancement algorithms. The root cause could be the disparities in the effectiveness of enhancement algorithms for feature extraction in analytic models. Specifically, the difference in class activation maps (CAMs) between enhanced and low-light frames demonstrates a positive correlation with video analytics accuracy. Motivated by such observations, a novel enhancement quality assessment method is proposed on CAMs to evaluate the effectiveness of different enhancement algorithms for low-light videos. Then, we design a multi-edge system, which adaptively offloads and enhances low-light video analytics tasks from mobile devices. To achieve the trade-off between the enhancement quality and the latency for all system-served mobile devices, we propose a genetic-based scheduling algorithm, which can find a near-optimal solution in a reasonable time to meet the latency requirement. Thereby, the offloading strategies and the enhancement algorithms are properly selected under the condition of limited end-edge bandwidth and edge computation resources. Simulation experiments demonstrate the superiority of the proposed system, improving accuracy up to 20.83\% compared to existing benchmarks.

Adaptive Offloading and Enhancement for Low-Light Video Analytics on Mobile Devices

TL;DR

A novel enhancement quality assessment method is proposed on CAMs to evaluate the effectiveness of different enhancement algorithms for low-light videos, and a genetic-based scheduling algorithm is proposed, which can find a near-optimal solution in a reasonable time to meet the latency requirement.

Abstract

In this paper, we explore adaptive offloading and enhancement strategies for video analytics tasks on computing-constrained mobile devices in low-light conditions. We observe that the accuracy of low-light video analytics varies from different enhancement algorithms. The root cause could be the disparities in the effectiveness of enhancement algorithms for feature extraction in analytic models. Specifically, the difference in class activation maps (CAMs) between enhanced and low-light frames demonstrates a positive correlation with video analytics accuracy. Motivated by such observations, a novel enhancement quality assessment method is proposed on CAMs to evaluate the effectiveness of different enhancement algorithms for low-light videos. Then, we design a multi-edge system, which adaptively offloads and enhances low-light video analytics tasks from mobile devices. To achieve the trade-off between the enhancement quality and the latency for all system-served mobile devices, we propose a genetic-based scheduling algorithm, which can find a near-optimal solution in a reasonable time to meet the latency requirement. Thereby, the offloading strategies and the enhancement algorithms are properly selected under the condition of limited end-edge bandwidth and edge computation resources. Simulation experiments demonstrate the superiority of the proposed system, improving accuracy up to 20.83\% compared to existing benchmarks.
Paper Structure (16 sections, 8 equations, 7 figures, 1 table, 1 algorithm)

This paper contains 16 sections, 8 equations, 7 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Motivation
  3. Detection of Different Low-light Videos
  4. Difference Between Enhancement and Low-Light
  5. System Model and Problem Formulation
  6. System Overview
  7. Enhancement Quality Assessment
  8. Problem Formulation
  9. Algorithm Design
  10. Performance Evaluation
  11. Experimental Settings
  12. Performance Comparison
  13. Impact of System Parameters
  14. The impact of bandwidth variation on accuracy
  15. The impact of the number of end devices on accuracy
  16. ...and 1 more sections

Figures (7)

  • Figure 1: Illustrations of example frames from two test videos.
  • Figure 2: Accuracy of the two videos with or without enhancement.
  • Figure 3: CAM comparisons of different frames.
  • Figure 4: CAM difference between enhanced and low-light frames and the accuracy of the enhanced frames.
  • Figure 5: System Model
  • ...and 2 more figures