Donald J. Patterson

Patent for infant gesture recognition

 video link

Patent Published!

It’s been a long time coming, but we finally had our patent formally published by the USPTO. Google has it listed under two numbers for some reason, US 9232912 B2 and US 20140066780 A1. The basic idea that we demonstrated was that if one appropriately monitors preterm babies then one can identify those that are susceptible to cerebral palsy and those that aren’t based on abnormal limb movements.   We used some machine learning and some techniques from ubiquitous computing to pull it off. The patent is just for the idea, but we also published a few papers demonstrating the efficacy.  We’ve celebrated all the milestones up to this point, so why not celebrate the last one as well!  One thing that I think is pretty funny is that I didn’t realize it was published until I started getting junk mail from companies that wanted to sell me a wide variety of trophies made out of the patent application – think little league photos for grown-ups.

System for evaluating infant movement using gesture recognition patent image

Custom Accelerometers

Another key aspect to our approach was that we used custom accelerometers to measure the babies. They were developed by Dr. Pai Chou, who was also at UCI at the time we did the work. They were very small, had a long battery life and were wireless: All characteristics that are helpful because they minimally impact the natural movement of the babies.  It would have been difficult to get clean data from devices that are heavy or had wires.

Limited study

Our study was somewhat limited in that we had a small population of babies that we monitored and they’re predisposed to developing brain hemorrhaging because the pregnancies were high risk.  This made our chance of finding problems higher than in an otherwise normal population, but there’s no reason why you couldn’t use it on a similar population of babies in the real world.

Next steps

So to follow up on this we are working on getting a bridge grant to increase the study size large enough that we can prove efficacy for FDA approval. Ideally a larger study will show the same results, we will find a company that wants to commercialize the technology and lots of babies will be helped by an early diagnosis. It’s kind of a long shot for all that to come together, “yet hope remains while the company is true”  Needless to say work like this is never done in a vacuum and I was fortunate to work with some great co-inventors to pull this off: Dana McQueen, Pai H. Chou, and Dan M. Cooper.


2 comments for “Patent for infant gesture recognition

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.