3D Face Recognition: How to make a fast and reliable database and compare the database with 2D+3D facial input?

摘要: A 3D face recognition algorithm has been developed for the Microsoft Kinect during final bachelor project at Delft University of Technology in 2013. The aim is to develop a prototype system. system outperform existing 2D main goal and divided into three parts. Each part was by group two students. subjects were data-acquisition, data-processing data comparison. Data-comparison topic this thesis. Nowadays, security an increasingly important society. can contribute make world safer place. This thesis about discovering new techniques chapter 4, but first getting familiar with 2D-world 3. Our results showed that not accurate use as too sensitive differences lightning, poses expressions. Meanwhile achieved our (proposed) quite fast accurate. had five correct matches from possible six matches. So only one person recognized process time 1.19 seconds. tested enough period created. It concluded works accurately geometry normal map input combination Haar-Walsh Transform angle-based distance, even though number sufficient yet. image input, both Haar transform distance. These options gave same result. future task we will increase confirm algorithm.