A study of methods of choosing the smoothing parameter in image restoration by regularization

Thompson, Andrew; Brown, J. C.; Kay, Jim; Titterington, D. M.

doi:10.1109/34.88568

Cited by 270 publications

(153 citation statements)

References 20 publications

Supporting

Mentioning

148

Contrasting

Order By: Relevance

“…However, choosing a proper regularization parameter is a key factor of the regularization method to obtain a good image estimate. Some popular methods for choosing the optimal regularization parameter have been developed in Thompson et al (1991) and Galatsanos and Katsaggelos (1992). Although the optimal regularization solution can lead to a perfect image estimate, only approximate optimal regularization parameters are available in practice since the noise variance needs to be estimated.…”

Section: Image Restoration and Estimationmentioning

confidence: 99%

Cooperative recurrent modular neural networks for constrained optimization: a survey of models and applications

Kamel

Xia

2008

Cogn Neurodyn

View full text Add to dashboard Cite

Constrained optimization problems arise in a wide variety of scientific and engineering applications. Since several single recurrent neural networks when applied to solve constrained optimization problems for real-time engineering applications have shown some limitations, cooperative recurrent neural network approaches have been developed to overcome drawbacks of these single recurrent neural networks. This paper surveys in details work on cooperative recurrent neural networks for solving constrained optimization problems and their engineering applications, and points out their standing models from viewpoint of both convergence to the optimal solution and model complexity. We provide examples and comparisons to shown advantages of these models in the given applications.

show abstract

Section: Image Restoration and Estimationmentioning

confidence: 99%

Cooperative recurrent modular neural networks for constrained optimization: a survey of models and applications

Kamel

Xia

2008

Cogn Neurodyn

View full text Add to dashboard Cite

show abstract

“…Our approaches differs from other methods based on residual statistics such as those in [8], [9] in those methods they don't use spectral properties of residual, but uses variance and other moments.BID suffers from an understanding lack of data. Very few assumptions are made on blurring filter and original image.…”

Section: International Journal For Research In Applied Science and Engimentioning

confidence: 86%

Blind and Non-Blind Deblurring using Residual Whiteness Measures

Rangaswamy¹

2017

IJRASET

View full text Add to dashboard Cite

In the present days blurring is a problem in the images and also in digital devices such as smart phone, digital camera, etc. Aim of the image deblurring is that it will make the pictures shape .In the existing method do not fing the perfect solutions, some disturbance are separately white that occurs in the image deblurring techniques .But in our proposed method when compared to the non blind deblurring and blind deblurring gives us the better results without noise both in single and multi frame scenarios and also evaluate the whiteness in the image in terms of speed and restoration quality when compared with the other deblurring techniques. This paper gives better results. Keywords: Image deploring, blind and Non-blind deploring, Whiteness in the image, multi frame scenarios. I.INTRODUCTION In 1960's peoples are concentrated more in producing pictures in the earth and the solar system. They need information from astronomical information so essentially need for the systematic image restoration techniques were what to the engineering community after solving in many algorithms solving extensively increase the need for astronomy and many other images they found a way called digital image deblurring .Generally images are classified into two types .They are constrained domain and unconstrained domain images. If there is no disturbance of light such images are referred as constrained images and if there is disturbance of light and posing problems in the are referred as by using camera but every image has more are less blur is occur due to a lot of interference in the camera are in the environment .Image deblurring is a general frame work is used to convert the measurements of the observed images into information about a physical object or system were the observed image is to make as a convolution of a applications in image deblurring are medical imaging, photography, surveillance. Mainly image deblurring is classified into two types they are blind and non-blind.If the blur Kernel is known then it is referred as non-blind deblurring and when blur Kernel and image are unknown than it is referred as blind deblurring. NBID has narrow applications when compared to BID due to ill-conditioned nature of blur operator. Most of the NBID methods overcome this problem with the help of image regularizer, or prior, the mean of which has to be tuned [1,[2]]. In ID several optimization techniques have been introduced to handle regularizers. Iterative shrinkage/thresholding (IST)algorithm is popular in handling regularizer. In these methods addition to regularizer parameter, choice of stopping criterion is also required, there is a delicate interplay between these two choices. Most BID methods restrict the use if blur filters, either in a hard way [3], [4] through the use of parametric models, or in a soft way [5], [6] through the use of regularizers. Recent BID method achieved good performance on synthetic and real problems without prior knowledge about blur. The above method works on estimating main features and regulariz...

show abstract

“…Approche relationnelle Rhenman (1964), Freeman (1984, Freeman & Red (1983), Bowie (1988), Alkhafaja (1989), Thompson et al (1991). Cornell & Shapiro (1987), Freeman & Evan (1990).…”

Section: Auteurs / Sourcesmentioning

confidence: 99%

Comment identifier les parties prenantes dans une certification RSE ?

Abboubi

2013

View full text Add to dashboard Cite

(imprimé) 1918-9222 (numérique) Découvrir la revue Citer cet article El Abboubi, M. (2013). Comment identifier les parties prenantes dans une certification RSE ?. Management international, 17(2), 48-62. doi:10.7202/1015400ar Résumé de l'article La présente recherche étudie l'identification des parties dans un processus de certification de Responsabilité Sociale de l'Entreprise (RSE) pour comprendre les mécanismes de sélection des parties prenantes prépondérantes. A travers une démarche qualitative basée sur deux études de cas, nous montrons que la certification RSE apporte des spécificités qui restent encore très peu documentées dans la littérature scientifique. Les parties prenantes identifiées dans les certifications RSE sont soit imposées par les fondements normatifs du standard, soit choisies par l'entreprise. L'identification est une opération de négociation et de construction tripartie dans laquelle l'entreprise, les parties prenantes et le certificateur peuvent jouer des rôles d'influence mutuels. L e management des parties prenantes constitue un défi managérial pour les entreprises engagées dans des processus de certifications liées à la responsabilité sociale (RSE). Celle-ci se définit comme un processus d'amélioration dans le cadre duquel les entreprises intègrent de manière volontaire, systémique et cohérente des considérations d'ordre social (Allouche, 2006). La mise en application de la démarche RSE via les certifications telles que le SA8000 (Social Accountability 8000), le global compact, le label social belge, le label égalité diversité ou encore la norme ISO 26000 soulève la question de la mobilisation des parties prenantes sans laquelle l'obtention de la certification ou du label est remise en cause (El abboubi, 2009). Certaines normes nomment clairement les parties prenantes à mobiliser dans le projet de certification, c'est le cas du label égalité diversité en Belgique. D'autres normes, comme le SA8000, le global compact ou encore l'ISO 26000 n'identifient que des catégo-ries de parties prenantes en utilisant un discours normatif qui laisse place à la subjectivité des managers dans leur prise de décision. Alors que la littérature scientifique dans ce domaine est riche et présente plusieurs variables détermi-nantes pour identifier les parties prenantes, elle passe cependant sous silence le caractère flexible et socialement construit des variables qui déterminent les parties prenantes lors d'un processus de certification RSE. Cette zone d'ombre motive la question de recherche adressée dans cette étude et pour laquelle nous avons apporté une réponse à travers une étude qualitative basée sur deux études de cas d'entreprises ayant choisi chacune une certification RSE déterminée. Une analyse restreinte au premier cas a révélé les étapes de mobilisation des parties prenantes de façon générale dans le processus de certification (références auteurs). La présente recherche entend focaliser la question de recherche sur l'identification des parties prenantes et ce en élargissant le terrain empi...

show abstract

A study of methods of choosing the smoothing parameter in image restoration by regularization

Cited by 270 publications

References 20 publications

Cooperative recurrent modular neural networks for constrained optimization: a survey of models and applications

Cooperative recurrent modular neural networks for constrained optimization: a survey of models and applications

Blind and Non-Blind Deblurring using Residual Whiteness Measures

Comment identifier les parties prenantes dans une certification RSE ?

Contact Info

Product

Resources

About