MBDFNet: Multi-scale Bidirectional Dynamic Feature Fusion Network for Efficient Image Deblurring

Zhongbao Yang,Jinshan Pan

DOI: https://doi.org/10.1109/ICME55011.2023.00096

2023-01-01

Abstract:Existing deep image deblurring models achieve favorable results with growing model complexity. However, these models cannot be applied to those low-power devices with resource constraints (e.g., smart phones) as these models usually have lots of networks parameters and require computational costs. To overcome this problem, we develop a multi-scale bidirectional dynamic feature fusion network (MBDFNet), a lightweight deep deblurring model, for efficient image deblurring, The proposed MBDFNet progressively restores multi-scale framework. To better utilize the features from coarse scale, we propose a bidirectional gated dynamic fusion module so that the most useful information of the features from coarse scales are kept to facilitate the estimations in the finer scale. We solve the proposed MBDFNet in an end-to-end FLOPs values, where the FLOPs value of the proposed MBDFNet is at least 6x smaller than the stat-of-the-art methods. Both MBDFNet achieves favorable performance in terms of model complexity while having competitive performance in terms of accuracy against stat-of-the-art methods.