Frequency-Driven Inverse Kernel Prediction for Single Image Defocus Deblurring
This work addresses the problem of recovering all-in-focus images from defocused ones for computer vision applications, representing an incremental improvement with novel frequency-domain integration.
The paper tackles the challenge of accurately modeling spatially varying blur kernels in single image defocus deblurring, where existing methods degrade in severely blurry regions, and proposes a Frequency-Driven Inverse Kernel Prediction network that incorporates frequency-domain representations to enhance kernel modeling, resulting in outperforming existing approaches in experiments.
Single image defocus deblurring aims to recover an all-in-focus image from a defocus counterpart, where accurately modeling spatially varying blur kernels remains a key challenge. Most existing methods rely on spatial features for kernel estimation, but their performance degrades in severely blurry regions where local high-frequency details are missing. To address this, we propose a Frequency-Driven Inverse Kernel Prediction network (FDIKP) that incorporates frequency-domain representations to enhance structural identifiability in kernel modeling. Given the superior discriminative capability of the frequency domain for blur modeling, we design a Dual-Branch Inverse Kernel Prediction (DIKP) strategy that improves the accuracy of kernel estimation while maintaining stability. Moreover, considering the limited number of predicted inverse kernels, we introduce a Position Adaptive Convolution (PAC) to enhance the adaptability of the deconvolution process. Finally, we propose a Dual-Domain Scale Recurrent Module (DSRM) to fuse deconvolution results and progressively improve deblurring quality from coarse to fine. Extensive experiments demonstrate that our method outperforms existing approaches. Code will be made publicly available.