Segregation and Context Aggregation Network for Real-time Cloud Segmentation

2025-04-19

AI Summary

To address the challenge of balancing accuracy and efficiency in ground-to-sky image cloud segmentation—particularly for deployment on resource-constrained edge devices—this paper proposes SCANet. Our method introduces a Separation and Context Aggregation Module (SCAM) that decouples sky and cloud features while enhancing both local and global contextual modeling. We adopt a lightweight CNN architecture and an ImageNet-free self-supervised pretraining strategy, further integrating FP16 inference optimization. Experiments demonstrate that SCANet-Large achieves state-of-the-art accuracy with a 70.9% reduction in parameter count compared to prior models; SCANet-Lite attains 1390 FPS on edge hardware, enabling ultra-real-time cloud segmentation. To the best of our knowledge, this is the first work to systematically reconcile accuracy, inference speed, and deployment efficiency for cloud segmentation—establishing a new benchmark for practical, edge-deployable solutions.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Systems and Infrastructure for Web, Mobile and WoT: Cloud, edge and content delivery systems for the Web
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Cloud segmentation from intensity images is a pivotal task in atmospheric science and computer vision, aiding weather forecasting and climate analysis. Ground-based sky/cloud segmentation extracts clouds from images for further feature analysis. Existing methods struggle to balance segmentation accuracy and computational efficiency, limiting real-world deployment on edge devices, so we introduce SCANet, a novel lightweight cloud segmentation model featuring Segregation and Context Aggregation Module (SCAM), which refines rough segmentation maps into weighted sky and cloud features processed separately. SCANet achieves state-of-the-art performance while drastically reducing computational complexity. SCANet-large (4.29M) achieves comparable accuracy to state-of-the-art methods with 70.9% fewer parameters. Meanwhile, SCANet-lite (90K) delivers 1390 fps in FP16, surpassing real-time standards. Additionally, we propose an efficient pre-training strategy that enhances performance even without ImageNet pre-training.

Problem

• Balancing cloud segmentation accuracy and computational efficiency
• Enabling real-time deployment on edge devices
• Reducing model parameters while maintaining performance

Innovation

• Lightweight SCANet model with SCAM module
• Separate processing of sky and cloud features
• Efficient pre-training strategy without ImageNet