Filters








977 Hits in 5.0 sec

Directly Invertible Nonlinear Divisive Normalization Pyramid for Image Representation [chapter]

Roberto Valerio, Eero P. Simoncelli, Rafael Navarro
2003 Lecture Notes in Computer Science  
We present a multiscale nonlinear image representation that permits an efficient coding of natural images.  ...  The parameters of the normalization operation are optimized in order to maximize the independence of the normalized responses for natural images.  ...  Image Representation Scheme The scheme proposed here consists of a linear decomposition followed by a nonlinear divisive normalization stage.  ... 
doi:10.1007/978-3-540-39798-4_42 fatcat:mnikrsqs3zgmhghk4fdn5fjafa

Statistically and perceptually motivated nonlinear image representation

Siwei Lyu, Eero P. Simoncelli, Bernice E. Rogowitz, Thrasyvoulos N. Pappas, Scott J. Daly
2007 Human Vision and Electronic Imaging XII  
We develop a reliable and efficient iterative procedure for inverting the divisive transformation.  ...  We describe an invertible nonlinear image transformation that is well-matched to the statistical properties of photographic images, as well as the perceptual sensitivity of the human visual system.  ...  We showed that this naturally leads to a nonlinear image representation based on a local divisive normalization transform.  ... 
doi:10.1117/12.720848 dblp:conf/hvei/LyuS07 fatcat:jloaqzxtqvhblpc73fh7z5crxe

Nonlinear image representation using divisive normalization

Siwei Lyu, Eero P. Simoncelli
2008 2008 IEEE Conference on Computer Vision and Pattern Recognition  
In this paper, we describe a nonlinear image representation based on divisive normalization that is designed to match the statistical properties of photographic images, as well as the perceptual sensitivity  ...  We further show that the resulting divisive normalization transform is invertible and provide an efficient iterative inversion algorithm.  ...  We developed methods for estimating the model parameters, and for inverting the divisive normalization transform.  ... 
doi:10.1109/cvpr.2008.4587821 pmid:25346590 pmcid:PMC4207373 dblp:conf/cvpr/LyuS08 fatcat:lwdh4ujayjha3ngwmnzntrm7au

End-to-end optimization of nonlinear transform codes for perceptual quality [article]

Johannes Ballé, Valero Laparra, Eero P. Simoncelli
2016 arXiv   pre-print
We introduce a general framework for end-to-end optimization of the rate--distortion performance of nonlinear transform codes assuming scalar quantization.  ...  When optimized over a large database of images, this representation offers substantial improvements in bitrate and perceptual appearance over fixed (DCT) codes, and over linear transform codes optimized  ...  Divisive normalization has previously been used in DCTbased image compression, e.g., [19, 20] .  ... 
arXiv:1607.05006v2 fatcat:proeznflajeqne4jwyx3zx4cmq

Nonlinear image representation for efficient perceptual coding

J. Malo, I. Epifanio, R. Navarro, E.P. Simoncelli
2006 IEEE Transactions on Image Processing  
We argue that linear transforms cannot achieve either of these goals and propose, instead, an adaptive nonlinear image representation in which each coefficient of a linear transform is divided by a weighted  ...  We then show that the divisive operation greatly reduces both the statistical and the perceptual redundancy amongst representation elements.  ...  INVERTING THE DIVISIVE NORMALIZATION TRANSFORM In order to use a normalized representation directly in a transform coding application, we need to be able to invert the transformation.  ... 
doi:10.1109/tip.2005.860325 pmid:16435537 fatcat:ck7sgb3cwzclvhozquh4zmapx4

Density Modeling of Images using a Generalized Normalization Transformation [article]

Johannes Ballé and Valero Laparra and Eero P. Simoncelli
2016 arXiv   pre-print
We introduce a parametric nonlinear transformation that is well-suited for Gaussianizing data from natural images.  ...  The transformation is differentiable and can be efficiently inverted, and thus induces a density model on images.  ...  DIVISIVE NORMALIZATION TRANSFORMATIONS Divisive normalization, a form of gain control in which responses are divided by pooled activity of neighbors, has become a standard model for describing the nonlinear  ... 
arXiv:1511.06281v4 fatcat:dz377i7icffm3hyppdgvjyiqbi

Efficient, nonlinear phase estimation with the nonmodulated pyramid wavefront sensor

Richard A. Frazin
2018 Optical Society of America. Journal A: Optics, Image Science, and Vision  
This article argues that nonlinear estimation based on Newton's method for nonlinear optimization can be useful for mitigating the effects of nonlinearity in the non-modulated PyWFS.  ...  The sensitivity of the the pyramid wavefront sensor (PyWFS) has made it a popular choice for astronomical adaptive optics (AAO) systems, and it is at its most sensitive when it is used without modulation  ...  Eq. (4) shows that the pyramid operator essentially takes two Fourier transforms of the pupil field (applying the phase ramp in between), thereby creating inverted images of the pupil.  ... 
doi:10.1364/josaa.35.000594 pmid:29603948 fatcat:iskabuoqhjennhmlxsywegcfjm

Spatio-chromatic information available from different neural layers via Gaussianization

Jesús Malo
2020 Journal of Mathematical Neuroscience  
Separate subsystems (e.g. opponent channels, spatial filters, nonlinearities of the texture sensors) have been suggested to be organized for optimal information transmission.  ...  How much visual information about the retinal images can be extracted from the different layers of the visual pathway?  ...  Acknowledgements I thank the attendees to my lectures on Information theory for Visual Neuroscience and particularly to Vir for encouraging me to write this with the promise of a travel.  ... 
doi:10.1186/s13408-020-00095-8 pmid:33175257 fatcat:gegtdfsxhzhw7k2mn2ci4miomq

Spatio-Chromatic Information available from different Neural Layers via Gaussianization [article]

Jesus Malo
2020 arXiv   pre-print
Separate subsystems (e.g. opponent channels, spatial filters, nonlinearities of the texture sensors) have been suggested to be organized for optimal information transmission.  ...  How much visual information about the retinal images can be extracted from the different layers of the visual pathway?.  ...  Acknowledgements: I thank the attendees to my lectures on Information theory for Visual Neuroscience and particularly to Vir for encouraging me to write this with the promise of a travel.  ... 
arXiv:1910.01559v3 fatcat:dvgtdcx6qfedvl7aodnipt45ba

FLIP-Q: A QCIF Resolution Focal-Plane Array for Low-Power Image Processing [chapter]

Jorge Fernández-Berni, Ricardo Carmona-Galán, Ángel Rodríguez-Vázquez
2012 Low-Power Smart Imagers for Vision-Enabled Sensor Networks  
The image processing primitives implemented by the chip, experimentally tested and fully functional, are scale space and Gaussian pyramid generation, fully-programmable multiresolution scene representation  ...  -including foveation -and block-wise energy-based scene representation.  ...  Pyramid representations solve this problem by subsampling the scale-space representations according to the filtering realized.  ... 
doi:10.1007/978-1-4614-2392-8_5 fatcat:dczuua54njglbbnwieo7wqhlca

FLIP-Q: A QCIF Resolution Focal-Plane Array for Low-Power Image Processing

Jorge Fernandez-Berni, Ricardo Carmona-Galan, Luis Carranza-Gonzalez
2011 IEEE Journal of Solid-State Circuits  
The image processing primitives implemented by the chip, experimentally tested and fully functional, are scale space and Gaussian pyramid generation, fully-programmable multiresolution scene representation  ...  -including foveation -and block-wise energy-based scene representation.  ...  Pyramid representations solve this problem by subsampling the scale-space representations according to the filtering realized.  ... 
doi:10.1109/jssc.2010.2102591 fatcat:f6tjgf5v5vfxbloy5xtmyehusm

What Do We Understand About Convolutional Networks? [article]

Isma Hadji, Richard P. Wildes
2018 arXiv   pre-print
In particular, spatial pyramid pooling is used to generate a fixed size representation independently from the size of the input image as illustrated in Figure 3 .13.  ...  Indeed, while early proposals of divisive normalization, e.g.  ... 
arXiv:1803.08834v1 fatcat:hhkitbh5x5gxtnss7y2pocgr7m

Cross Dissolve Without Cross Fade: Preserving Contrast, Color and Salience in Image Compositing

Mark Grundland, Rahul Vohra, Gareth P. Williams, Neil A. Dodgson
2006 Computer graphics forum (Print)  
It is a full time occupation for many artists working with digital media.  ...  image compo-  ...  Salience preserving pyramid blending uses standard pyramid blending with weights specified by salience mattes calculated at each level from Gaussian pyramid representations of the component images (the  ... 
doi:10.1111/j.1467-8659.2006.00977.x fatcat:p2wi2nortzg2xnt53ingooofca

End-to-end Optimized Image Compression [article]

Johannes Ballé, Valero Laparra, Eero P. Simoncelli
2017 arXiv   pre-print
We describe an image compression method, consisting of a nonlinear analysis transformation, a uniform quantizer, and a nonlinear synthesis transformation.  ...  Using a variant of stochastic gradient descent, we jointly optimize the entire model for rate-distortion performance over a database of training images, introducing a continuous proxy for the discontinuous  ...  ACKNOWLEDGMENTS We thank Olivier Hénaff and Matthias Bethge for fruitful discussions.  ... 
arXiv:1611.01704v3 fatcat:h4gk2xthhrhffay6qy46dc55uq

Survey of Texture Mapping

Paul S. Heckbert
1986 IEEE Computer Graphics and Applications  
Williams uses a box filter to construct the image pyramid, but gaussian filters can also be E used [Bur81] . He also proposes a particular layout for color image pyramids called the ''mipmap''.  ...  techniques for parameterization, scaning, texture representation, direct convolution, and prefiltering.  ... 
doi:10.1109/mcg.1986.276672 fatcat:vgxzsxnuerhkpgg3hdeblyu7ay
« Previous Showing results 1 — 15 out of 977 results