I am a PhD student at the Department of Computer Science Engineering, Michigan State University. I am working under the supervision of Prof. Anil K. Jain in the Pattern Recognition and Image Processing (PRIP) Lab. My research interests include applying computer vision and machine learning techniques in fingerprint biometrics.
Antecedent to joining PRIP Lab, I worked at IBM Research Lab, India, where my work focused on automating the process of capturing customer name variants in IBM sales and services data using data mining and approximate text matching techniques.
I graduated with a Bachelors of Technology (Hons.) degree in Computer Science and Engineering (2013) from IIIT Delhi.
For a detailed view of my work, please download my resume or refer my profile and projects section.
August 2015 - Present
August 2009 - December 2013
Department Rank: 2 (w/ Hons.)
(i) Image Analysis and Machine Intelligence
(ii) Data Analytics
(iii) Mobile Computing
August, 2015 - Present
May, 2018 - August, 2018
February, 2014 - August 2015
December, 2013 - February, 2014
May, 2013 - July, 2013
May, 2012 - November, 2012
Designated duties included designing and conducting lab sessions, doubt clearing sessions, office hours, checking quizzes and examinations.
Designed an 8-week long online refresher module on C Language majorly for incoming UnderGrad and Grad students. Designed the course work, homework assignments, quizzes and a programming competition in conjunction with a team of 3 junior TAs.
Responsible for designing the content of weekly lab sessions and mentoring students to conduct lab sessions for fellow students under my supervision. This rotational activity gave every student a chance to prepare content as a teacher resulting in better understanding.
I have witnessed my 5 minutes of fame.
T. Chugh, K. Cao, and Anil K. Jain, Fingerprint Spoof Buster: Use of Minutiae-centered Patches, IEEE Transactions on Information Forensics and Security (TIFS), Vol. 13, No. 9, pp. 2190-2202, DOI: 10.1109/TIFS.2018.2812193, 2018 [Video]
T. Chugh, K. Cao, J. Zhou, E. Tabassi and A. K. Jain, Latent Fingerprint Value Prediction: Crowd-based Learning, IEEE Transactions on Information Forensics and Security (TIFS), Vol. 13, No. 1, pp. 20-34, DOI 10.1109/TIFS.2017.2721099, Jan 2018 [Invited Talk @ IAI, 2016] [poster] [Crowdsourcing Tool] [Nist.gov] [ Ishi News]
T. Chugh and A. K. Jain, Fingerprint Presentation Attack Detection: Generalization and Efficiency, in 12th IAPR International Conference on Biometrics (ICB), Crete, Greece, 2019 [spotlight] [poster]
R. Gajawada, A. Popli, T. Chugh, A. Namboodiri, and A. K. Jain, Universal Material Translator: Towards Spoof Fingerprint Generalization, in 12th IAPR International Conference on Biometrics (ICB), Crete, Greece, 2019
E. Tabassi, T. Chugh, D. Deb, A. K. Jain, Altered Fingerprints: Detection and Localization, in 9th IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS), Los Angeles, October, 2018 [poster]
T. Chugh, K. Cao, A. K. Jain, Fingerprint Spoof Detection Using Minutiae-based Local Patches, IEEE International Joint Conference on Biometrics (IJCB), Denver, Colorado, October, 2017 [spotlight] [poster]
T. Chugh, S. S. Arora, A. K. Jain, and N. G. Paulter, Benchmarking Fingerprint Minutiae Extractors, in 16th IEEE International Conference of the Biometrics Special Interest Group (BIOSIG), Darmstadt, Germany, September, 2017 [oral]
T. Chugh, M. Singh, S. Nagpal, R. Singh, and M. Vatsa, Transfer Learning Based Evolutionary Algorithm for Composite Face Sketch Recognition, in IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Honolulu, Hawaii, July, 2017 [oral]
T. Chugh, H.S. Bhatt, R. Singh, and M. Vatsa, Matching Age Separated Composite Sketches and Digital Face Images, in 6th IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS), Washington D.C., October, 2013 [oral]
T. Chugh and A. K. Jain, OCT Fingerprints: Resilience to Presentation Attacks, arXiv:1908.00102, 2019
D. Deb, T. Chugh, J. Engelsma, K. Cao, N. Nain, J. Kendall, A. K. Jain, Matching Fingerphotos to Slap Fingerprint Images, arXiv:1804.08122, 2018
T. Chugh, K. Cao, J. Zhou, E. Tabassi and A. K. Jain, Latent Fingerprint Value Prediction: Crowd-based Learning, MSU Technical Report, MSU-CSE-16-4, 2016
T. Chugh, Bridging the two BitTorrent Worlds I2P and P2P, INRIA Technical Report, November, 2012
Fingerprint Spoof Buster Lite utilizes quantized tensorflow-lite MobileNet CNN models for efficient fingerprint spoof detection on a commodity smartphone. It evaluates local patches centered around minutiae clusters and averages the patch scores to produce the global spoofness score. Additionally, it also highlights the live and spoof regions in the input fingerprint image. The fingerprint image can be loaded from phone memory or as a live-scan captured by a fingerprint reader connected to the phone via OTG cable.Android Application
FingerprintMash, inspired from FaceMash, is a crowdsourcing tool designed to collect quality labels and pairwise comparisons between pairs of latent fingerprint images from a crowd of fingerprint experts. The collected labels were utilized to identify the underlying bases used by fingerprint experts to determine value of a given latent, and consequently, learn an objective latent value predictor in terms of automatically extracted latent features.Crowdsourcing Tool
A deep-learning based approach for real-time automatic tattoo detection and identification to assist law enforcement agencies. This work was done during my summer internship at NEC Labs America, Princeton in 2018.Tattoo Recognition
This tool ranks law and safety events, detected on twitter, based on their veracity and impact. It uses user/event level spatio-temporal features and incremental onlineSVM for ranking. Current work includes implementing module for rumor detection. In future, this project will be included in a dashboard designed to assist law enforcement agencies.Event Detection, Machine Learning
This tool is designed to eliminate customer name discrepancies (missing/incomplete information) in IBM sales and services data. It links all name variants to a single identifier. Using "approximate text-matching" techniques, this tool scores and ranks the detected name variants.Text-Matching, Industrial Research
This project utilizes a transfer learning approach to match composite sketches and digital face images. The knowledge is transferred from information rich source domains. By using an evolutionary algorithm, the parameters are learnt in source domains and transferred using inductive transfer learning techniques. It also eliminates the need for re-training.Composite Sketches, Transfer Learning
This project aims to assist law-enforcement agencies in matching composite sketches with digital face images. A state-of-the-art algorithm, specifically designed to perform when there is an age gap between composite sketches and digital images, is presented in the work published at BTAS’13 (Oral).Face Recognition, Research
This project included designing an interface to visualize power consumption in IIIT Delhi campus. Used mango automation for designing interface and connected it to SQL server database. The interface was fed by live data from Raspberry Pi sensor systems placed at 40+ electricity meters.Mango Automation, Energy Dashboard
This food-ordering android application provides a key-feature of “live inventory”. Gone are those days when you would need to call and ask the vendor about the availability of your relished dish. This app displays the available quantity and gives you freedom of cashless food ordering. Also a boon to vendors, now their inventory is well managed.Android Application, Website
Designed a bridge to interconnect anonymous file sharing network “i2p” and content-rich BitTorrent network. Added features like DHT support, Cache Manager, UDP Tracker Support and Multi Bridge Protocol. Work acknowledged in the research paper.I2P, BitTorrent, P2P, INRIA
This application considers user location, budget and food preferences to recommend restaurants. The suggestions are budget friendly and located equi-distant from all members of the group. It is powered by Google App Engine, GeoLocation API and Foursquare API.MeetUp, Recommender System, Fair Policy
Landscape and portrait photography excites me.Photography, Website
I like to strum my guitar.Music, Guitar
I like to cycle on long trails. Maximum distance covered is 110km in 2 day routine in Colmar, France.Cycling
Beginner in speed cubing. Best time: 82 secondsCube