Signal processing, machine learning, and video tell the truth

Description

Technology has finally caught up to—and surpassed—our ability to lie. We can now discover (and sometimes even predict) a person’s emotions and intent, making it easy to determine whether you have evil inclinations… or are simply coming down with the flu.

David Maman discusses the critical role of signal processing in modern data science applications. Backed by centuries of proven mathematical applications, when signal processing is applied to data preparation, it accelerates delivery of the right answers to critical business questions. Signal processing eliminates noise and prepares data for machine learning and artificial intelligence, significantly reducing false positives and delivering actionable results. When data is denoised correctly, it eliminates unnecessary components and leaves only good data to be modeled and analyzed, raising the accuracy of analysis to previously unattainable levels.

David demonstrates the efficacy of signal processing and machine learning using video and remote photo plethysmography (PPG) heart-rate detection and the transformation to heart-rate variability (HRV), leads a live demo, and shares an analysis of famous people, like President Donald Trump.

PPG is a noninvasive, low-cost, simple optical technique that detects blood volume changes in the microvascular bed of tissue. After the recording is complete, the PPG information is transformed into HRV using multiple denoising mechanisms.

Take the raw data from the video, the HRV results, and some facial expression recognition, put them through basic signal processing, add in some finely tuned algorithms, and use machine and deep learning with a bit of AI thrown in to get a clear picture.

While the common people are sure that autonomous vehicles are going to ignore the health condition of the driver and passengers, the automotive vendors understand that it’s the exact opposite. The passengers’ health conditions become a high priority in the age of autonomous vehicles to prevent cases where a passenger will arrive deceased to his destination. HRV has been shown to correlate to risk of sudden death from cerebrovascular accident (CVA), cognitive function, insulin sensitivity, inflammatory markers, visceral fat, aerobic fitness levels, fatigue markers and overtraining, performance, and others.