Efficient Transformer-based Visible to Infrared Image Translation Model for Enhanced Downstream Applications

The proposed end-to-end Transformer-based model efficiently translates visible light images into high-fidelity infrared images by leveraging a Dynamic Fusion Aggregation Module and an Enhanced Perception Attention Module to capture and preserve crucial textural and color features.