Edge-Based Defocus Blur Estimation With Adaptive Scale Selection

Karaali, Ali; Jung, Cláudio Rosito

doi:10.1109/tip.2017.2771563

Cited by 108 publications

(121 citation statements)

References 34 publications

Supporting

Mentioning

121

Contrasting

Order By: Relevance

“…Since the pioneering work of Zhuo and Sim (2011), many edgebased methods (Tang et al, 2013;Chen et al, 2016;Liu et al, 2016;Karaali and Jung, 2018) rely on gradient magnitude ratios, obtained by reblurring the input image and computing the amount of defocus blur from the ratio between the gradients of the input and those reblurred images. This and many other sparse defocus estimators assume an isotropic 2D Gaussian PSF, although an optical or system's PSF is never truly Gaussian.…”

Section: Edge-based Methodsmentioning

confidence: 99%

“…Most existing techniques apply a slow Laplacian-based interpolation scheme, which makes dense defocus map extraction very time-consuming for images with large pixel counts. Faster approaches relying on superpixels (Chen et al, 2016), the fast guided filter (Andrade, 2016;Karaali and Jung, 2018) or sparse blur map downsampling (Kriener et al, 2013) have recently been introduced as well.…”

Section: Edge-based Methodsmentioning

confidence: 99%

“…More recently, Karaali combined the pixel-and transform-based methods since they both explore image patches. He contrasted these patch-based approaches with edge-based methods (Karaali and Jung, 2018).…”

Section: Single Image Defocus Mappingmentioning

confidence: 99%

“…Since it is not feasible to report on all possible methods in the context of this paper, only three state-of-the-art but completely different edge-based methods will be assessed: Andrade 2016; Chen et al (2016); Karaali and Jung (2018). Despite the power of some recent patch-based methods, edge-based approaches remain a very attractive choice.…”

Section: Defocus Mapping Toolboxmentioning

confidence: 99%

See 3 more Smart Citations

Focusing on Out-of-Focus: Assessing Defocus Estimation Algorithms for the Benefit of Automated Image Masking

Verhoeven

2018

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Acquiring photographs as input for an image-based modelling pipeline is less trivial than often assumed. Photographs should be correctly exposed, cover the subject sufficiently from all possible angles, have the required spatial resolution, be devoid of any motion blur, exhibit accurate focus and feature an adequate depth of field. The last four characteristics all determine the “sharpness” of an image and the photogrammetric, computer vision and hybrid photogrammetric computer vision communities all assume that the object to be modelled is depicted “acceptably” sharp throughout the whole image collection. Although none of these three fields has ever properly quantified “acceptably sharp”, it is more or less standard practice to mask those image portions that appear to be unsharp due to the limited depth of field around the plane of focus (whether this means blurry object parts or completely out-of-focus backgrounds). This paper will assess how well- or ill-suited defocus estimating algorithms are for automatically masking a series of photographs, since this could speed up modelling pipelines with many hundreds or thousands of photographs. To that end, the paper uses five different real-world datasets and compares the output of three state-of-the-art edge-based defocus estimators. Afterwards, critical comments and plans for the future finalise this paper.

show abstract

Section: Edge-based Methodsmentioning

confidence: 99%

Section: Edge-based Methodsmentioning

confidence: 99%

Section: Single Image Defocus Mappingmentioning

confidence: 99%

Section: Defocus Mapping Toolboxmentioning

confidence: 99%

See 2 more Smart Citations

Focusing on Out-of-Focus: Assessing Defocus Estimation Algorithms for the Benefit of Automated Image Masking

Verhoeven

2018

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…Single image defocus blur estimation techniques mainly categorized into two classes, ie, gradient‐based methods and frequency‐based methods . The stated classes are further narrowed into two subclasses: the first one is edge‐based methods and the second one is region‐based method . In edge‐based methods, blur estimation is performed at edge locations.…”

Section: Introductionmentioning

confidence: 99%

Single image defocus estimation by modified gaussian function

Hassan

Abbasi

Waqas

et al. 2019

Trans Emerging Tel Tech

View full text Add to dashboard Cite

This article presents an algorithm to estimate the defocus blur from a single image. Most of the existing methods estimate the defocus blur at edge locations, which further involves the reblurring process. For this purpose, existing methods use the traditional Gaussian function in the phase of reblurring but it is found that the traditional Gaussian kernel is sensitive to the edges and can cause loss of edges information. Hence, there are more chances of missing spatially varying blur at edge locations. We offer the repeated averaging filters as an alternative to the traditional Gaussian function, which is more effective, and estimate the spatially varying defocus blur at edge locations. By using repeated averaging filters, a blur sparse map is computed. The obtained sparse map is propagated by integration of superpixels segmentation and transductive inference to estimate full defocus blur map. Our adopted method of repeated averaging filters has less computational time of defocus blur map estimation and has better visual estimates of the final defocus recovered map. Moreover, it has surpassed many previous state‐of‐the‐art proposed systems in terms of quantative analysis.

show abstract