Every day it gets a little bit easier to make the phone call from inside the house.
AI-driven video analyzer sets new standards in human action detection
Oct 2024, phys.org
If you were one person in a sea of thousands, we can now tell you're up to something by the way you move your body, with only one camera:
Semantic and Motion-Aware Spatiotemporal Transformer Network - a multi-feature selective attention model helps the system focus on the most important parts of a scene while ignoring unnecessary details, and a motion-aware 2D positional encoding algorithm tracks how things move over time, to accurately recognize complex actions in real time, making it more effective in high-stakes scenarios like surveillance, health care diagnostics, or autonomous driving.
via University of Virginia's School of Engineering and Applied Science: Matthew Korban et al, A Semantic and Motion-Aware Spatiotemporal Transformer Network for Action Detection, IEEE Transactions on Pattern Analysis and Machine Intelligence (2024). DOI: 10.1109/TPAMI.2024.3377192
Image credit: Colonial algae Volvox spheres in a drop of water by Dr. Jan Rosenboom - Nikon Small World Photomicrography Competition - 2025
A stapler that knows when you need it: Using AI to turn everyday objects into proactive assistants
Oct 2025, phys.org
And there you have it, the AI stapler.
via Carnegie Mellon University: Violet Yinuo Han et al, Towards Unobtrusive Physical AI: Augmenting Everyday Objects with Intelligence and Robotic Movement for Proactive Assistance, Proceedings of the 38th Annual ACM Symposium on User Interface Software and Technology (2025). DOI: 10.1145/3746059.3747726
New 'hidden in plain sight' facial and eye biomarkers for tinnitus severity could unlock path to testing treatments
Apr 2025, phys.org
Researchers knew that the pupil dilation was a sign of increased arousal and that involuntary facial movements could provide a window into threat assessment; and they hypothesized that people with debilitating tinnitus are chronically in vigilance mode, reacting to everyday sounds as if they are threats.Video recordings were made while participants listened to pleasant, neutral, or distressing and unpleasant sounds. In people with severe tinnitus, pupils dilated extra wide at all sounds, while facial movements were blunted in response to the same sounds.
via Mass General Brigham Eye and Ear: Samuel Smith et al, Objective autonomic signatures of tinnitus and sound sensitivity disorders, Science Translational Medicine (2025). DOI: 10.1126/scitranslmed.adp1934.
Deepfakes now come with a realistic heartbeat, making them harder to unmask
Apr 2025, phys.org
Photoplethysmography - The analysis of the transmission of light through the skin and underlying blood vessels has long been indispensable in medicine, for example in pulse oximeters. Its digital cousin, so-called remote photoplethysmography (rPPP), is an emerging method in telehealthcare, which uses webcams to estimate vital signs. But rPPP can, in theory, also be used in deepfake detectors.In recent years, such experimental rPPP-based deepfake detectors have proven good at distinguishing between real and deepfaked videos. These successes led some experts to judge that current deepfakes couldn't yet mimic a realistic heart rate. But now, it appears that this complacent view is outdated."Our experiments have shown that current deepfakes may show a realistic heartbeat, but do not show physiologically realistic variations in blood flow across space and time within the face."
via Humboldt University of Berlin: High-quality deepfakes have a heart!, Frontiers in Imaging (2025). DOI: 10.3389/fimag.2025.1504551
 |
Nikon Small World - Sunflower pollen on an acupuncture needle - John-Oliver Dum - 40X - 2023 |
Diagnostic pen converts handwriting into electrical signals to detect Parkinson's
Jun 2025, phys.org
The pen's design integrates a silicone-based magnetoelastic tip embedded with magnetic particles, and a reservoir of ferrofluid ink containing nanomagnets. During writing, mechanical stress from hand pressure deforms the tip and causes ink movement, generating shifts in magnetic flux that induce voltage in an embedded coil.
via UCLA: Guorui Chen et al, Neural network-assisted personalized handwriting analysis for Parkinson's disease diagnostics, Nature Chemical Engineering (2025). DOI: 10.1038/s44286-025-00219-5
Also: Diagnosing Parkinson's disease using a soft magnetoelastic smart pen, Nature Chemical Engineering (2025). DOI: 10.1038/s44286-025-00228-4
One of a kind: Humans have unique breathing 'fingerprints' that may signal health status
Jun 2025, phys.org
Scientists can identify individuals based solely on their breathing patterns with 96.8% accuracy.This high-level accuracy remained consistent across multiple retests conducted over a two-year period, rivaling the precision of some voice recognition technologies. Respiratory fingerprints correlated with a person's body mass index, sleep-wake cycle, levels of depression and anxiety, and even behavioral traits. For example, participants who scored relatively higher on anxiety questionnaires had shorter inhales and more variability in the pauses between breaths during sleep.
via Noam Sobel of the Weizmann Institute of Science: Humans Have Nasal Respiratory Fingerprints, Current Biology (2025). DOI: 10.1016/j.cub.2025.05.008.
Banking data reveals early warning signs of cognitive decline in older adults
Jun 2025, phys.org
The results reveal that subtle but significant changes in financial behavior—such as reduced spending on travel and hobbies, increased household bills, fewer online banking logins, and more frequent requests to reset PINs — begin to appear several years before individuals are formally identified as lacking financial capacity.
"Lacking financial capacity" is another way of saying it.
via University of Nottingham's School of Economics: Anna Trendl et al, Early Behavioral Markers of Loss of Financial Capacity, JAMA Network Open (2025). DOI: 10.1001/jamanetworkopen.2025.15894
WhoFi: New surveillance technology can track people by how they disrupt Wi-Fi signals
Jul 2025, phys.org
Total and absolute surveillance is inevitable.
WhoFi captures how Wi-Fi signals change when they encounter people and objects, interpreting unique Wi-Fi disruptions as individual "fingerprints" unique to a person.The team acknowledges the issues and potential for misuse, but states that their re-identification (Re-ID) system does not capture a person's identity or personal data. "By leveraging non-visual biometric features embedded in Wi-Fi CSI, this study offers a privacy-preserving and robust approach for Wi-Fi-based Re-ID, and it lays the foundation for future work in wireless biometric sensing."
via Sapienza University: Danilo Avola et al, WhoFi: Deep Person Re-Identification via Wi-Fi Channel Signal Encoding, arXiv (2025). DOI: 10.48550/arxiv.2507.12869
No comments:
Post a Comment