Data Augmentation for Improving SSD Performance in Rainy Weather Conditions

Quoc-Viet Hoang, Trung-Hieu Le, Shih-Chia Huang

Abstract: This paper focuses on using data augmentation to improve the performance of Single Shot Multibox Detector (SSD) in rainy weather conditions. Data augmentation is applied by generating synthetic rain on a set of clear images to expand the dataset for training. The experimental results show that the performance of SSD model improves by up to 16.82% when being trained on the proposed augmented synthetic dataset.