Congratulations to Jian Li for passing his defense on November 29, 2016. His Ph.D. thesis is entitled “Advanced Techniques for Latent Fingerprint Enhancement and Recognition”. Here is an abstract of his thesis and we are so glad to have him share his Ph.D. experience as well.
Fingerprints provide one of the most popular biometric data, and have been widely used in individual person identification and verification. The Automated Fingerprint Identification System (AFIS) offers important evidences for criminal investigation, and serves as an important tool for law enforcement. As compared with conventional exemplar fingerprints, latent fingerprints are typically collected in a crime scene. They are often degraded and corrupted with very poor quality, leading to very low identification rates. In a practical system, a latent fingerprint has to be enhanced prior to feature extraction to ensure a reliable fingerprint matching performance. In this research, we study techniques for latent fingerprint enhancement and orientation field estimation to achieve a higher matching rate. Our studies include traditional image processing techniques as well as a new method based on the emerging convolutional neural network (CNN).
First, we propose a new method using the Markov random field (MRF) model and the sparse representation (SR) of ridges to enhance latent fingerprint. The proposed MRF-SR method is inspired by the recent success of dictionary-based methodologies in recent fingerprint community. Second, we extend the MRF-SR method in order to extract the orientation field of fingerprint. As an essential feature of fingerprints, the orientation field can enhance a fingerprint image with directional and contextual filtering. It provides a valuable supplementary tool to other orientation estimation algorithms in the literature. Finally, we explore the feasibility and study the performance of applying the CNN to latent fingerprint enhancement as our main focusing in this talk. Being motivated by recent developments of CNN in image enhancement and restoration applications, we propose a novel encoding-decoding neural network, called the FingerNet. The FingerNet is trained in a pixelwise end-to-end manner for direct fingerprint enhancement. In particular, we develop a novel data augmentation method to add structured noise to good quality fingerprint patches so as to form meaningful training data. Then, we design a multi-task encoder-decoder network that has a leading convolutional module, which is followed by two deconvolutional branches. We successfully train the proposed FingerNet and it outperforms benchmarking methods with a fast computational speed.
It is really a great pleasure and honour for me to pursue the PhD degree at MCL family. Research topics in our group are always interesting, with a combination of theory and practice. With Prof. Kuo’s guidance, we can learn the most three important characters in my opinion: good taste, thorough thinking, and strong implementation. Besides, during this process, we need to get along well with ourselves to keep positive attitude, enough patience and strong confident. Indeed, this procedure will definitely make us better in terms of both intelligence and emotional quotient. Additionally, there are a lot of interesting and practical projects, that can help us strengthen our professional knowledge and skills. In one word, it will be a treasure if you treasure every chance you can improve.
I will join Apple after my graduation and start a new journey of my career. I believe that the PhD experience will help me a lot for the new position in industry. I want to contribute what I leaned at our group to Apple so that my value can be realized, maybe in the next Apple products. Also, I will keep a close connection with our group if I can provide any help in the future.
Congratulations again to Jian and we wish him all the best for his career in Apple.