AQIFormer: A Transformer-Based Multi-View Architecture for Cross-City Air Quality Classification
Researchers have developed AQIFormer, a new transformer-based architecture designed to classify air quality using images. This model integrates front and rear traffic imagery with weather data, improving cross-city generalization and achieving 89.96% accuracy on a large dataset. AQIFormer demonstrates strong performance even on unseen cities, with minimal accuracy degradation when adapted with few-shot learning. AI
IMPACT This model offers a more scalable and cost-effective approach to air quality monitoring, potentially improving environmental health insights.