Pattern-based palm recognition for identification system using deep learning
THESIS STUDENTS:
Simon Mahdavi
From: Uppsala, Sweden
Studies: Biomedical Engineering with a Master’s in Image analysis, Data Analysis, Signal Processing, and Machine Learning at Lund University
Leo Li
From: Linköping, Sweden
Studies: Biomedical Engineering with a Master’s in Image analysis, Data Analysis, Signal Processing, and Machine Learning at Lund University
How did you become aware of Precise Biometrics, and what attracted you to do your project with us?
Leo: I had heard of Precise Biometrics before, but Simon and I had not considered the company for our thesis until we saw the LinkedIn ad. Since both of us specialize in Image Analysis and Machine Learning, we thought that Precise would be an excellent place to apply our skills in a real-world setting and learn from industry experts. During our meeting with Kerstin Johnsson and Anders Olsson from Precise, we were impressed by the company’s culture and overall atmosphere. Plus, the view from the office isn’t too shabby either!
What are the main research questions you aim to answer through your thesis?
- Using Machine Learning to quantify and categorize hand posture, for the purpose of improving region of interest segmentation.
- User friendly system design for an interactive palm identification system.
Leo: Our primary research question is centered around measuring hand posture and the degree of openness of the palm’s surface. This poses a significant challenge in obtaining a clear and accurate image that can be effectively matched by the matching algorithm. Therefore, our focus is on developing effective methods to measure and analyze these specific aspects of the hand to enhance the accuracy and reliability of our image matching algorithm.
How do you believe your thesis contributes to the existing knowledge in AI and biometric technology?
Simon: Biometric technology has been evolving rapidly in recent years and is now seen in various devices and services in our everyday life. I really believe that the unique features of an individual’s palm can be of importance in the future for biometrics and we have already seen companies like Amazon and Apple using the advantages of palm recognition. Hopefully, some of the work performed in this thesis can help Precise adding a palm recognition system to the product portfolio.
In what ways do you believe your research could be expanded or improved upon in the future?
Simon: There is a lot of work left to do within the area. Security aspects of a palm recognition systems must be explored, both identifying risks and further development on the matcher algorithm. When compared to face recognition or fingerprint recognition, a palm recognition system requires more work on the design aspects and user experience, since it is contactless, and the posture can vary in every user case.
Leo: A significant portion of our work involves developing a methodology for extracting the Region of Interest (ROI)* and evaluating hand posture, as well as creating a user-friendly design for obtaining the best possible image for matching. I believe that these methods can serve as a blueprint for future research in similar implementations.
What are the practical implications of your research findings, and how do you envision them being applied in real-world settings?
Leo: Palm recognition offers several advantages, including contactless operation, which makes it a more hygienic solution suitable for large settings such as hospitals. In addition to hygiene, palm recognition provides enhanced security and privacy. Our research seeks to address some of the challenges associated with palm recognition and contributes to expanding the range of biometric solutions available for widespread use. By exploring this technology, we hope to create a more secure and reliable system for various applications.
GLOSSARY*
ROI – Region of Interest – refers to a specific area or feature of a biometric sample that is of particular interest for analysis or comparison.