Rabbi Dr. Ari Berman, President and Rosh Yeshiva | Yeshiva University
Rabbi Dr. Ari Berman, President and Rosh Yeshiva | Yeshiva University
A team of researchers, including Dr. Yucheng Xie from the Katz School's Graduate Computer Science and Engineering department, has introduced a new technology named "mmPalm." This innovation uses millimeter wave signals for palm recognition to provide a low-effort authentication method aimed at securing environments such as hotel entryways, apartment buildings, and vehicles.
The system is detailed in the paper titled “mmPalm: Unlocking Ubiquitous User Authentication Through Palm Recognition with mmWave Signals.” Unlike traditional biometric systems that depend on fingerprints or facial recognition, mmPalm offers a cost-effective alternative that could transform security access in public spaces like smart cities and homes.
The study was awarded the Best Paper Follow-Ups award at the IEEE Conference on Communications and Network Security (CNS), which gathers leading researchers and industry experts globally to discuss advancements in cybersecurity.
Traditional biometric methods often require costly hardware for installation and maintenance. In contrast, mmPalm leverages mmWave technology used in WiGig and 5G networks to identify individuals based on their unique palm patterns. Dr. Xie stated, “The technology is promising because of the fine detail that mmWave can capture.”
The system captures these characteristics by sending and analyzing reflected signals to create a distinctive "palm print" for each user. It also builds virtual antennas to enhance spatial resolution further, capturing subtle differences in each palm print.
For users, authentication involves simply showing their hand. The device transmits frequency-modulated waves interacting with the palm; mmPalm then analyzes these reflected waves for specific traits. This data forms a unique biometric profile compared against stored profiles for identity verification.
Beyond being cost-effective, mmPalm addresses challenges like distance and hand orientation using artificial intelligence called Conditional Generative Adversarial Network (cGAN) to generate virtual profiles. A transfer learning framework allows adaptation to different environments.
Testing with 30 participants over six months demonstrated a 99% accuracy rate with high resistance to impersonation and spoofing. As smart technology becomes more prevalent in cities and homes, mmPalm could lead the way for secure contactless user authentication.
“By harnessing high-resolution mmWave signals to extract detailed palm characteristics,” said Dr. Xie, “mmPalm presents an ubiquitous, convenient and cost-efficient option to meet the growing needs for secure access in a smart, interconnected world.”