Efficient Decoder and Intermediate Domain for Semantic Segmentation in Adverse Conditions
-
Published:2024-01-19
Issue:1
Volume:7
Page:254-276
-
ISSN:2624-6511
-
Container-title:Smart Cities
-
language:en
-
Short-container-title:Smart Cities
Author:
Chen Xiaodong1, Jiang Nan1, Li Yifeng2, Cheng Guangliang2, Liang Zheng3, Ying Zuobin4, Zhang Qi4, Zhao Runsheng5
Affiliation:
1. School of Statistics, Renmin University of China, Beijing 100872, China 2. School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou 510275, China 3. Celedyne Technologies Co., Ltd., Shenzhen 518057, China 4. Faculty of Data Science, City University of Macau, Macau SAR 999078, China 5. National Instruments Corporation, Shanghai 201210, China
Abstract
In smart city contexts, traditional methods for semantic segmentation are affected by adverse conditions, such as rain, fog, or darkness. One challenge is the limited availability of semantic segmentation datasets, specifically for autonomous driving in adverse conditions, and the high cost of labeling such datasets. To address this problem, unsupervised domain adaptation (UDA) is commonly employed. In UDA, the source domain contains data from good weather conditions, while the target domain contains data from adverse weather conditions. The Adverse Conditions Dataset with Correspondences (ACDC) provides reference images taken at different times but in the same location, which can serve as an intermediate domain, offering additional semantic information. In this study, we introduce a method that leverages both the intermediate domain and frequency information to improve semantic segmentation in smart city environments. Specifically, we extract the region with the largest difference in standard deviation and entropy values from the reference image as the intermediate domain. Secondly, we introduce the Fourier Exponential Decreasing Sampling (FEDS) algorithm to facilitate more reasonable learning of frequency domain information. Finally, we design an efficient decoder network that outperforms the DAFormer network by reducing network parameters by 28.00%. When compared to the DAFormer work, our proposed approach demonstrates significant performance improvements, increasing by 6.77%, 5.34%, 6.36%, and 5.93% in mean Intersection over Union (mIoU) for Cityscapes to ACDC night, foggy, rainy, and snowy, respectively.
Subject
Electrical and Electronic Engineering,Artificial Intelligence,Urban Studies
Reference59 articles.
1. Karmouni, H., Jahid, T., El Affar, I., Sayyouri, M., Hmimid, A., Qjidaa, H., and Rezzouk, A. (2017, January 22–24). Image analysis using separable Krawtchouk-Tchebichef’s moments. Proceedings of the 2017 International Conference on Advanced Technologies for Signal and Image Processing (ATSIP), Fez, Morocco. 2. Avazov, K., Mukhiddinov, M., Makhmudov, F., and Cho, Y.I. (2021). Fire detection method in smart city environments using a deep-learning-based approach. Electronics, 11. 3. Image classification using separable invariant moments of Charlier-Meixner and support vector machine;Hmimid;Multimed. Tools Appl.,2018 4. A Hybrid Algorithm for Urban LULC Change Detection for Building Smart-city by Using WorldView Images;Pal;IETE J. Res.,2023 5. Jahid, T., Karmouni, H., Hmimid, A., Sayyouri, M., and Qjidaa, H. (2017, January 15–18). Image moments and reconstruction by Krawtchouk via Clenshaw’s reccurence formula. Proceedings of the 2017 International Conference on Electrical and Information Technologies (ICEIT), Rabat, Moroccan.
|
|