Paper Image

Adapting object detection to nighttime driving

Published on:

2 April 2024

Primary Category:

Computer Vision and Pattern Recognition

Paper Authors:

Jicheng Yuan,

Anh Le-Tuan,

Manfred Hauswirth,

Danh Le-Phuoc

Bullets

Key Details

Employs global-local transformations to reduce domain gap

Leverages proxy student network for oversight avoidance

Introduces adaptive thresholding to expand target search space

Achieves state-of-the-art performance on real and synthetic datasets

AI generated summary

Adapting object detection to nighttime driving

This paper proposes a Cooperative Students framework to adapt object detection models from daytime to nighttime driving conditions. It uses global-local transformations and proxy student networks to capture spatial consistency across domains and avoid overlooking potential detections. Comprehensive experiments show superior performance over prior state-of-the-art techniques.

Answers from this paper

Comments

No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up