In the summer of 2019, I had the opportunity to work as a research intern at Indian Institute of Technology, Indore under the guidance of Dr. Somnath Dey, Assistant professor and the then acting head of Discipline of Computer Science and Engineering.
Together with a fellow intern, my work included research on Masterprint and ways to prevent it and implementing an algorithm for robust matching of partial fingerprints using minutiae and ridge shape features.
After the initial research work, my first task was to familiarize myself with NIST Biometric Image Software (NBIS) and process the FVC databases with it. Afterwards, I created partial fingerprints using the FVC databases and running the Bozorth3 matching algorithm on it for benchmarking. This results would then be compared to the to-be-made algorithm. A huge number of partial images were generated from the FVC database, therefore, an automated system was requred, it was achieved using Python, from generating the partial fingerprints to running Bozorth3 algorithm on them.
After the reference result was collected, the work on the algorithm implementaion began, it was wrtten in MATLAB because of heavy image processing tasks and computation. After 3-4 weeks, the implementation was ready and it was run on the partial fingerprints and the results were collected for processing. The results compared on a small set of images showed promsing results. As of writing this post, the results are yet to be compared on the full database because of the computation time and size of the partial fingerprint database. Once the data is processed and compared, the results will be shown seperately in much more detail.
It was an amazing learning experience, both from academic and non-academic point of view. I learned MATLAB, got better in using Linux, learned a lot about fingerprints and biometrics in general and the importance of documentation. I also learned to work better in a team and to rebound from a bad day at work and try to make the next one better.