LHN etV2: A Balanced L ow-cost H ybrid N etwork for Single Image Dehazing

Shenghai Yuan,Jijia Chen,Wenchao Jiang,Zhiming Zhao,Song Guo

DOI: https://doi.org/10.1109/tmm.2024.3377133

IF: 7.3

2024-01-01

IEEE Transactions on Multimedia

Abstract:Single-image dehazing is a challenging task that requires both local details and global distribution. Existing methods face challenges in color imbalance and inconsistent details when predicting a haze-free image, because of their limitations in generalization from a specific setting (physics-based methods), capturing global information (CNN-based methods) and capturing detailed local information ( ViT-based methods). In response to these challenges, we propose a balanced low-cost hybrid network called LHNetV2 based on LHNetV1. The key insight of LHNetV2 is the effective fusion of different features, and a series of novel approaches is proposed to increase the running speed of the original LHNetV1. Firstly, building upon the Feature-aware Information Fusion method, we preserve the original Physical Embedding and Architecture Aggregation components in LHNetV1. Next, to overcome the speed bottleneck of LHNetV1, we enhance the calculation method of attention in the ViT sub-network and streamline the cross-stage interaction strategy in the CNN main-network. Finally, we introduce a dynamic adversarial loss function to bolster both the training stability and performance of LHNetV2. The experiments are extensively conducted on mainstream datasets, and the results demonstrate that LHNetV2 achieves the best balance between the performance and the running speed in single-image dehazing. The code is available at https://github.com/SHYuanBest/LHNet.

computer science, information systems,telecommunications, software engineering