Light Invariant Video Imaging for Improved Performance of Convolution Neural Networks

Amir Kolaman, Dan Malowany, Rami Hagege, Hugo Guterman

Abstract:

Light conditions affect the performance of computer vision
algorithms by creating spatial changes in color and
intensity across a scene. Convolutional neural networks
(CNN) use color components of the input image and as
a result are sensitive to ambient light conditions. This
work analyzes the influence of ambient light conditions on
CNN classifiers. We suggest a method to boost the performance
of CNN-based object detection and classification
algorithms by using Light Invariant Video Imaging (LIVI).
LIVI neutralizes the influence of ambient light conditions
and renders the perceived object’s appearance independent
of the light conditions. Training sets consists mainly, if not
only, of objects in natural light conditions. As such, using
LIVI boosts CNN performance by matching object appearance
to that expected by the CNN model, which was created
according to the training set. We further investigate
the use of LIVI as a feedback source for a robust ongoing
re-training mechanism for CNN. Faster Region-based CNN
(FRCNN) was used as a case study in order to validate the
importance of light conditions on CNN performance and on
how it can be improved by using LIVI as an input or as a
feedback source for the re-training process. We show that
LIVI enables reduced CNN size, enhanced performance,
and improved training.