Publication Date

Spring 2016

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science

First Advisor

Robert Chun

Second Advisor

Thomas Austin

Third Advisor

Amit Singh Jadon


Facial Recognition Neural Nets GPU programming


In field of computer vision research, One of the most important branch is Face recognition. It targets at finding size and location of human face on digital image, by identifying and separating faces from the surrounding objects like building, plants etc. For the purpose of developing an advanced face recognition algorithm, Detection of facial key points is the basic and very important task, basically it is about finding out the location of specific key points on facial images. This key points can be mouths, noses, left eyes, right eyes and so on.

For implementation of solution, I have used amazon ec2 gpu instance and convolutional networks consisting of multiple levels. Outputs of multiple networks are fused at every level for accurate and robust evaluation. At the stage of initialization, high level features are extracted over the whole face region which helps in locating key points with high accuracy. Local minimum occurred by data corruption and ambiguity in difficult samples of image caused by occlusions, extreme lightings and large variations in poses can be avoided by this method. At later levels, training of networks is adopted to locally refine the initial predictions and the input supplied to them are limited to smaller regions around the predictions that are obtained in initial stage.