Color Representation in Deep Neural Networks

Convolutional neural networks are top-performers on image classification tasks. Understanding how they make use of color information in images may be useful for various tasks. In this paper we analyze the representation learned by a popular CNN to detect and characterize color-related features. We confirm the existence of some object- and color-specific units, as well as the effect of layer-depth on color-sensitivity and class-invariance.


Presented at:
IEEE International Conference on Image Processing (ICIP), Beijing, China, September 17-20, 2017
Year:
2017
Laboratories:




 Record created 2017-10-12, last modified 2018-03-17

n/a:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)