Real-world Scene Image Enhancement with Contrastive Domain Adaptation Learning

Abstract

Image enhancement methods leveraging learning-based approaches have demonstrated impressive results when trained on synthetic degraded-clear image pairs. However, when deployed in real-world scenarios, such models often suffer significant performance degradation due to the inherent domain gap between synthetic and real degradations. To bridge this gap, we propose a novel T wo-stage C ontrastive D omain A daptation image E nhancement framework ( TCDAE ) consisting of two key strategies: 1) S ynthetic-to- R eal D omain T ransfer L earning ( S2R-DTL ) that effectively translates images from the synthetic degraded domain to the real degraded domain, aligning the domains at the pixel level, and 2) D egraded-to- C lear D omain T ransfer L earning ( D2C-DTL ) that further adapts the enhancement model from the synthetic to the real domain by translating images from the real degraded domain to the real clean domain in both supervised and unsupervised branches. A unique aspect of our approach is the integration of a Do main N oise C ontrastive E stimation ( DoNCE ) loss in both learning strategies. This specialized loss formulation enables TCDAE to robustly translate images across domains, even in scenarios lacking strong positive examples. Consequently, our framework can generate enhanced images with natural, realistic appearances akin to real clear images. Comprehensive experiments on real-world degraded scenes across diverse tasks including dehazing, deraining, and deblurring demonstrate the superiority of TCDAE over state-of-the-art methods, achieving improved visual quality, quantitative metrics, and downstream task performance.