Menu
Află starea comenzii
Contactați-ne
Carte Support Vector Machine Learning Jonathan Robinson

Support Vector Machine Learning

Application to Compression of Digital Images

Autor: Jonathan Robinson
Limbă: engleză
Legare: Carte broșată
Editura: VDM Verlag
Disponibilitate: În depozitul extern
Expediem în 9-15 zile
378.61 lei
Methods exploring the application of support vector§machine learning (SVM) to still image compressio...

Informații despre carte

Autor
Jonathan Robinson
Limbă
engleză
Legare
Carte - Carte broșată
Publicat
2008
Pagini
176
EAN
9783639100006
ISBN
363910000X
Enbook ID
06819008
Editura
VDM Verlag
Greutate
245
Dimensiuni
152 x 229 x 10

Descriere completă

Methods exploring the application of support vector§machine learning (SVM) to still image compression are§detailed in both the spatial and frequency domains.§In particular the sparse properties of SVM learning§are exploited in the compression algorithms. A§classic radial basis function neural network requires§that the topology of the network be defined before§training. An SVM has the property that it will choose§the minimum number of training points to use as§centres of the Gaussian kernel functions. It is this§property that is exploited as the basis for image§compression algorithms presented in this book.§Several novel algorithms are developed applying SVM§learning to both directly model the colour surface§and model transform coefficients after the surface§has been transformed into the frequency domain. It is§demonstrated that compression is more efficient in§frequency space.§In the frequency domain, results are superior to that§of JPEG. For example, the quality of the industry§standard Lena image compressed 63:1 for JPEG is§slightly worse quality than the same image compressed§192:1 with the RKi-1 algorithm detailed in this book. Methods exploring the application of support vector§machine learning (SVM) to still image compression are§detailed in both the spatial and frequency domains.§In particular the sparse properties of SVM learning§are exploited in the compression algorithms. A§classic radial basis function neural network requires§that the topology of the network be defined before§training. An SVM has the property that it will choose§the minimum number of training points to use as§centres of the Gaussian kernel functions. It is this§property that is exploited as the basis for image§compression algorithms presented in this book.§Several novel algorithms are developed applying SVM§learning to both directly model the colour surface§and model transform coefficients after the surface§has been transformed into the frequency domain. It is§demonstrated that compression is more efficient in§frequency space.§In the frequency domain, results are superior to that§of JPEG. For example, the quality of the industry§standard Lena image compressed 63:1 for JPEG is§slightly worse quality than the same image compressed§192:1 with the RKi-1 algorithm detailed in this book.

S-ar putea să te intereseze

Trans-Allegheny Pioneers (West Virginia and Ohio)

John P Hale
151.58 lei

Lotos-Time in Japan. ... Illustrated.

Henry Theophilus Finck
144.00 lei

Once Upon a Zzzz

Maddie Frost
73.51 lei

Dialogue on Kabbalah by Samuel David Luzzatto

Menachem Kellner
257.56 lei

Atlantic Cataclysm

David Eltis
165.13 lei

Here Lies a Lovable Lawyer

Chris Rogers
229.95 lei

Clienții care au cumpărat această carte au mai cumpărat și

Gesetz zu dem Vertrag vom 14. April 2014 zwischen der Bundesrepublik Deutschland und der Weltgemeinschaft Reformierter Kirchen ? Körperschaft des öffe

Antiphon Verlag
45.90 lei

Nyuma ya Mapazia

R. Mtobwa
97.27 lei