Rain Rendering and Construction of IRain Vehicle Color/I-24 Dataset

The fine identification of vehicle color can assist in criminal investigation or intelligent traffic management law enforcement. Since almost all vehicle-color datasets that are used to train models are collected in good weather, the existing vehicle-color recognition algorithms typically show poor...

Full description

Saved in:
Bibliographic Details
Published in:Mathematics (Basel) 2022-09, Vol.10 (17)
Main Authors: Hu, Mingdi, Wang, Chenrui, Yang, Jingbing, Wu, Yi, Fan, Jiulun, Jing, Bingyi
Format: Article
Language:English
Subjects:
Machine vision
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The fine identification of vehicle color can assist in criminal investigation or intelligent traffic management law enforcement. Since almost all vehicle-color datasets that are used to train models are collected in good weather, the existing vehicle-color recognition algorithms typically show poor performance for outdoor visual tasks. In this paper we construct a new Rain Vehicle Color-24 dataset by rain-image rendering using PS technology and a SyRaGAN algorithm based on the Vehicle Color-24 dataset. The dataset contains a total of 40,300 rain images with 125 different rain patterns, which can be used to train deep neural networks for specific vehicle-color recognition tasks. Experiments show that the vehicle-color recognition algorithms trained on the new dataset Rain Vehicle Color-24 improve accuracy to around 72% and 90% on rainy and sunny days, respectively. The code is available at humingdi2005@github.com.
ISSN:2227-7390
2227-7390
DOI:10.3390/math10173210