The purpose of this diploma thesis is to create a mathematical program,
which would convert simple pictures of faces into a 3D model. The program
would be used for forensic applications, where it would enable an easier
detection of faces. With a 3D model, identification would be more accurate,
unlike identification from a simple image or video. The program would be
created in the Python environment, where the it would use the image as an
input, where the image would be stripped for data for the program to attempt
to recreate an as accurate as possible 3D model of the face in the image. The
method, which the program would use is depth detection, where the depth
would be converted to a 3D model. With the help of neural networks we
were able to successfully create a copy of a face in a 3D environment.
|
