Vortex Fourier encoding for small-brain classification of MNIST digits with no hidden layers

Baurzhan Muminov; Luat T. Vuong

doi:10.1117/12.2558983

20 May 2020 Vortex Fourier encoding for small-brain classification of MNIST digits with no hidden layers

Baurzhan Muminov, Luat T. Vuong

Proceedings Volume 11388, Image Sensing Technologies: Materials, Devices, Systems, and Applications VII; 113880T (2020) https://doi.org/10.1117/12.2558983
Event: SPIE Defense + Commercial Sensing, 2020, Online Only

Abstract

Here we elaborate on the edge-enhanced spectral components that are produced by the vortex Fourier trans- form, which are introduced in [1]. The vortex phase pattern imprinted on from an object breaks the spatial invariance of its Fourier representation is robust to noise. We report on new results related to the image classification of the MNIST digit dataset with no hidden layers. We show that the accuracy from one phase vortex mask is capable of achieving 0:95 validation accuracy and further show that the dynamic range of the phase modulation scheme significantly influences the classification accuracy and classification convergence rate.

Citation Download Citation

Baurzhan Muminov and Luat T. Vuong "Vortex Fourier encoding for small-brain classification of MNIST digits with no hidden layers", Proc. SPIE 11388, Image Sensing Technologies: Materials, Devices, Systems, and Applications VII, 113880T (20 May 2020); https://doi.org/10.1117/12.2558983

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available