Deakin University
Browse

File(s) under permanent embargo

A convolutional neural network for pixelwise illuminant recovery in colour and spectral images

conference contribution
posted on 2018-01-01, 00:00 authored by Antonio Robles-KellyAntonio Robles-Kelly, Ran Wei
Here, we present a pixelwise illuminant recovery method for both, trichromatic and multi or hyperspectral images which employs a convolutional neural nettwork. The network used here is based upon the simple, yet effective architecture employed by the CIFARIO-quick net[1]. The network is trained using a loss function which employs the angular difference between the target illuminant and the estimated one as the data term. The loss used here also includes a regularisation term which encourages smoothness in the spectral domain. Moreover, the network takes, at input, a tensor which is constructed making use of an image patch at different scales. This allows the network to predict the illuminant per-pixel using locally supported multiscale information. We illustrate the utility of our method for both, colour and hyperspectal illuminant recovery and compare our results against other techniques elsewhere in literature.

History

Event

Pattern Recognition. International Conference (24th : 2018 : Beijing, China)

Pagination

109 - 114

Publisher

IEEE

Location

Beijing, China

Place of publication

Piscataway, N.J.

Start date

2018-08-20

End date

2018-08-24

ISSN

1051-4651

ISBN-13

9781538637883

Language

eng

Publication classification

E Conference publication; E1 Full written paper - refereed

Title of proceedings

ICPR 2018 : Proceedings of the 24th International Conference on Pattern Recognition