Learning to Deblur using Light Field Generated and Real Defocus Images

Ruan, Lingyan; Chen, Bin; Li, Jizhou; Lam, Miu Ling

doi:10.1109/cvpr52688.2022.01582

Cited by 37 publications

(32 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…IFANet [32] adopts extra dual views during the training. Table 1 demonstrates that our method substantially outperforms existing CNNbased methods, improving over the state-of-the-art method DRBNet [53] by 0.68dB (+2.7%) PSNR, and slightly outperforms Transformer-based methods, yielding +0.17dB improvement over Restormer [84]. Notably, our method requires much less computational effort, saving up to 32.2% parameters and 39% MACs.…”

Section: Comparison To the State-of-the-art Methodsmentioning

confidence: 85%

“…1), where only for Restormer [84] the SSIM metric shows slightly better results than our method. Additionally, we follow the training strategy that initiates the training on the LFDOF dataset [54] followed by the DPDD dataset, showing a better generalization ability than the existing state-of-the-art DRBNet [53], as illustrated in Tab. 3 and Fig.…”

Section: Comparison To the State-of-the-art Methodsmentioning

confidence: 99%

“…However, these two-step strategies are less effective in terms of quality and computational cost due to the error propagation occurring in their iterative optimization procedure. Later, CNN-based methods predominate in the image deblurring task by offering end-to-end solutions [2,11,28,32,43,48,53,63,85,86], and efficiently achieving remarkable results, most of which are tailored to a specific type of blur -motion or defocus. Transformer-based structures (e.g.…”

Section: Related Work Image Deblurringmentioning

confidence: 99%

“…These two properties can potentially be quantified using the estimated parameters of our fitted f (x). Specifically, [32], KPAC [63], GKM [48], Restormer [84], MDP [1], DRBNet [53]. Note that DRBNet in this case is trained using only the DPDD dataset.…”

Section: Erfmetermentioning

confidence: 99%

“…In the last two decades, CNNs have become a promising tool for image deblurring tasks. Given a large dataset, CNNs have the ability to learn the corresponding priors, which then can be used for image deblurring at inference [2,11,28,32,43,53,63,85,86], showing high efficiency and generalization ability [57]. While on the one hand, the inherent inductive biases contribute to the efficiency of the network; on the other hand, it limits the network's ability to model long-range spatial dependencies.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Revisiting Image Deblurring with an Efficient ConvNet

Ruan¹,

Bemana²,

Seidel³

et al. 2023

Preprint

View full text Add to dashboard Cite

Image deblurring aims to recover the latent sharp image from its blurry counterpart and has a wide range of applications in computer vision. The Convolution Neural Networks (CNNs) have performed well in this domain for many years, and until recently an alternative network architecture, namely Transformer, has demonstrated even stronger performance. One can attribute its superiority to the multi-head self-attention (MHSA) mechanism, which offers a larger receptive field and better input content adaptability than CNNs. However, as MHSA demands high computational costs that grow quadratically with respect to the input resolution, it becomes impractical for high-resolution image deblurring tasks. In this work, we propose a unified lightweight CNN network that features a large effective receptive field (ERF) and demonstrates comparable or even better performance than Transformers while bearing less computational costs. Our key design is an efficient CNN block dubbed LaKD, equipped with a large kernel depth-wise convolution and spatial-channel mixing structure, attaining comparable or larger ERF than Transformers but with a smaller parameter scale. Specifically, we achieve +0.17dB / +0.43dB PSNR over the state-of-theart Restormer on defocus / motion deblurring benchmark datasets with 32% fewer parameters and 39% fewer MACs. Extensive experiments demonstrate the superior performance of our network and the effectiveness of each module. Furthermore, we propose a compact and intuitive ERFMeter metric that quantitatively characterizes ERF, and shows a high correlation to the network performance. We hope this work can inspire the research community to further explore the pros and cons of CNN and Transformer architectures beyond image deblurring tasks.

show abstract

Section: Comparison To the State-of-the-art Methodsmentioning

confidence: 85%