ISNAS-DIP: Image-Specific Neural Architecture Search for Deep Image Prior

Arican, Metin Ersin; Kara, Özgür; Bredell, Gustav; Konukoğlu, Ender

doi:10.1109/cvpr52688.2022.00200

Cited by 10 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[13] restricts the search space to the network components related to the up-sampling part and modifies the original architecture by considering cross-scale residual connections, i.e., the skip connections link blocks of the encoder part with blocks of the decoder part at different spatial dimensions. In [14] the focus is shifted on the metrics used to evaluate the performance: the authors develop several novel metrics that allow to reduce the costs of NAS procedures. It is worth to notice that, differently from the approach proposed in this paper, in none of the previously cited works on NAS for DIP, a performance predictor has been employed to improve the DIP behaviour.…”

Section: Introductionmentioning

confidence: 99%

Early stopping strategies in Deep Image Prior

Benfenati,

Catozzi,

Franchini

et al. 2024

Preprint

View full text Add to dashboard Cite

The so-called Deep Image Prior approach is an unsupervised deep learning methodology which has gained great interest in recent years due to its effectiveness in tackling imaging problems. However, a well known drawback of Deep Image Prior is the need for a proper early stopping technique to prevent undesired corruptions in the reconstructed images. Determining the optimal number of iterations depends on both the specific application and the features of the images being processed. As a consequence, several numerical trials are typically required to decide when to stop the Deep Image Prior procedure, resulting in significant computational costs and time requirements. This paper aims to introduce two early stopping techniques for Deep Image Prior, based on different approaches and different perspectives. The first one relies on the Neural Architecture Search strategy by generating hyperparameters configurations for the neural network employed in Deep Image Prior, configurations that shall be able to provide clean images comparable to those obtained by the standard configuration, optimally stopped, but with significantly fewer iterations. The second proposed early stopping strategy is based on a modified version of the BRISQUE metric, a no-reference image quality measure, and it aims to track the behaviour of the PSNR curve, obtained by applying Deep Image Prior, without knowing the ground truth image. While the NAS-based early stopping technique is particularly suited in those situations where the computational time is limited, this latter one is also relevant when a larger number of iterations is allowed. Several numerical experiments on different denoising applications show a promising performance of Deep Image Prior combined with the suggested early stopping procedures.

show abstract

Section: Introductionmentioning

confidence: 99%

Early stopping strategies in Deep Image Prior

Benfenati,

Catozzi,

Franchini

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Later, Mataev et al [30] enhanced the DIP framework by introducing a technique called 'regularized denoising' (RED) as a priori. In another study, Ersin Arican et al [31] proposed a neural structure search (NAS) strategy specifically designed for image processing within the DIP framework. This strategy requires less training compared to conventional NAS methods and effectively implements image-specific NAS, thereby reducing the search space and improving efficiency.…”

Section: Introductionmentioning

confidence: 99%

Combining Deep Image Prior and Second-Order Generalized Total Variance for Image Inpainting

You

Yang

et al. 2023

Mathematics

View full text Add to dashboard Cite

Image inpainting is a crucial task in computer vision that aims to restore missing and occluded parts of damaged images. Deep-learning-based image inpainting methods have gained popularity in recent research. One such method is the deep image prior, which is unsupervised and does not require a large number of training samples. However, the deep image prior method often encounters overfitting problems, resulting in blurred image edges. In contrast, the second-order total generalized variation can effectively protect the image edge information. In this paper, we propose a novel image restoration model that combines the strengths of both the deep image prior and the second-order total generalized variation. Our model aims to better preserve the edges of the image structure. To effectively solve the optimization problem, we employ the augmented Lagrangian method and the alternating direction method of the multiplier. Numerical experiments show that the proposed method can repair images more effectively, retain more image details, and achieve higher performance than some recent methods in terms of peak signal-to-noise ratio and structural similarity.

show abstract

Advanced Image Processing Techniques for Ultrasound Images using Multiscale Self Attention CNN

Vetriselvi,

Thenmozhi

2023

Neural Process Lett

View full text Add to dashboard Cite

ISNAS-DIP: Image-Specific Neural Architecture Search for Deep Image Prior

Cited by 10 publications

References 10 publications

Early stopping strategies in Deep Image Prior

Early stopping strategies in Deep Image Prior

Combining Deep Image Prior and Second-Order Generalized Total Variance for Image Inpainting

Advanced Image Processing Techniques for Ultrasound Images using Multiscale Self Attention CNN

Contact Info

Product

Resources

About