2021
DOI: 10.1016/j.knosys.2020.106617
|View full text |Cite
|
Sign up to set email alerts
|

A deep learning based image enhancement approach for autonomous driving at night

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
60
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
7
3

Relationship

1
9

Authors

Journals

citations
Cited by 160 publications
(60 citation statements)
references
References 59 publications
0
60
0
Order By: Relevance
“…These advancements have not only been in the areas of physiological sensing and wearable devices. Over the past decade, there have been significant improvements in computer vision and machine learning approaches, where we can now accurately detect specific features and behaviors of drivers from the in-cabin videos while detecting objects [36] and outside conditions through the outdoor videos [37]. However, since the majority of existing NDS were introduced over a decade ago, many of the existing datasets do not include these modalities of data such as driver's pose features, gaze patterns, and objects in the environment.…”
Section: Introductionmentioning
confidence: 99%
“…These advancements have not only been in the areas of physiological sensing and wearable devices. Over the past decade, there have been significant improvements in computer vision and machine learning approaches, where we can now accurately detect specific features and behaviors of drivers from the in-cabin videos while detecting objects [36] and outside conditions through the outdoor videos [37]. However, since the majority of existing NDS were introduced over a decade ago, many of the existing datasets do not include these modalities of data such as driver's pose features, gaze patterns, and objects in the environment.…”
Section: Introductionmentioning
confidence: 99%
“…However, the proposed methods are based on insulator images captured in daytime, while detecting insulator defects at night is still a challenging task. Therefore, one of our future work will focus on improving the night detection performance by using image enhancement technologies [51]. Moreover, researchers have introduced the concept of Corner Net [52] and Extreme Net [53] into target detection in the recent years, and satisfactory results have been reported.…”
Section: Discussionmentioning
confidence: 99%
“…More specifically, the positive detection of a target can always trigger the tracking of a specific object. In other words, the proposed approach may not work well at night or some cases of insufficient lighting(hazy weather) [49]- [51]. Second, the proposed approach effectively fits the 3D vehicle model into the image frame when the vehicle is tracked, but the vehicle model fitting evolutionary procedure is time-consuming.…”
Section: E Discussionmentioning
confidence: 99%