Abstract:
Varying illumination conditions have always been bottleneck in recognizing faces under
wild environment. Because changing the lightning conditions and position alter the shadows,'
brightness and shadows which result in the form of changing facial features, which as a whole
can turn the facial geometry and therefore made it difficult for a system to detect a face under
such variations.
Our goal here was to develop a system which can recognize face, unaffected of the
illumination invariance and Convolutional Neural Networks allow us to extract a wide range of
features from images. Turns out, we can use this idea of feature extraction for face recognition
too under wild illumination conditions.
i!
ii
11
ii
:
The approach we are going to use for face recognition is fairly straight forward. The key
here is to get a deep neural network to produce a bunch of numbers that describe a face (known
as face encodings). When you pass different images of the same person under different
illumination impact, the network should return similar outputs (i.e. closer numbers) foi both
images, whereas when you pass in the images of two different people, the netwoik should return
'
;
very different outputs for the two images. This means that the neuial netwoik needs to be tiained
on that. The ; to automatically identify different features of faces and calculate numbeis based
output of the neural network can be thought of as an identifier for a paiticular peison s face — if
you pass in different images of the same person, the output of the neural network will be very
similar/close, whereas if you pass in images of a different person, the output will be very,
different and all this process is running under varying illumination conditions and proves that
control these variations and give fruitful results independent of the
:
;
CNNs are good enough to
way in which lightning is to be used.
