My US Patent — In Simple Words
Assessing the severity of respiratory distress with a 3D camera. Small sentences, easy to say.
What this patent is
I am a named inventor on a US patent that came out of my PhD.
It protects a way to measure how severe a child's breathing distress is, using a 3D camera.
So my research became real, protected technology — not just a paper.
Who is on the patent
I am the first inventor: Haythem Rehouma.
With me: Prof. Rita Noumeir (my PhD supervisor).
And two medical doctors: Dr. Philippe Jouvet and Dr. Sandrine Essouri, from Sainte-Justine hospital.
So it joins engineering and medicine on the same invention.
More detail — if they ask
Filed as a PCT international application on Sep 23, 2020, from a US provisional filed Sep 24, 2019. The owner is SOCOVAR, the company that manages intellectual property for ÉTS — so the university handled the filing, which is normal for academic inventions.
The core idea, in one breath
A 3D camera looks at the chest and the abdomen of the patient.
A computer finds a point on the chest and a point on the abdomen.
It measures the distance between them, and compares it to a threshold.
From that, it gives a signal of how severe the breathing distress is — with no contact.
Why it deserved a patent
Today, doctors judge chest retractions and see-saw breathing with the eyes — subjective.
There is no standard medical device that puts a number on chest-wall retraction.
My invention gives that objective number, automatically, from a camera.
That gap is exactly what makes it novel and useful.
The main claim, step by step
1. Use a 3D camera to make an image of the chest and abdomen.
2. Find the chest point and the abdomen point in that image.
3. Compute the thoraco-abdominal distance between them.
4. Compare it to a threshold, and output a severity signal.
More detail — if they ask
This is the independent claim — the broad protection. The dependent claims then add specifics: using two cameras, matching point clouds, surface reconstruction, computing tidal volume and respiratory rate, and detecting thoraco-abdominal asynchrony (TAA).
Finding inspiration and expiration
The patent also protects a motion-over-time method.
I follow a reference point on the chest-abdomen surface across many frames.
When the motion changes direction, that is the moment breathing switches.
So I can mark the end of inhale and the end of exhale, automatically.
How the system works
Two cameras give raw depth → two point clouds.
I match the clouds, then segment the chest-abdomen region.
I rebuild the 3D surface, then compute the volume.
From the volume over time, I get the breathing parameters.
More detail — if they ask
Point clouds are cleaned with a Statistical Outlier Removal filter, normals oriented with a minimum spanning tree, and the surface built with Poisson reconstruction. Sensors used include the Kinect v2 and the Microsoft Azure DK, both time-of-flight RGB-D cameras.
What the invention measures
1. The respiratory rate — how fast.
2. The tidal volume — air per breath.
3. The minute ventilation — air per minute.
4. The retraction distance and the see-saw motion (TAA).
Does it work? Yes.
I tested on a high-fidelity baby mannequin and against a laser distance sensor.
I simulated four modes: normal, mild, severe, irregular.
The camera and the laser agreed very well — correlation above 0.985.
I also studied the best camera position around the bed for the ICU.
More detail — if they ask
Retraction distances were about 1.95 mm (mild), 3.64 mm (severe), 2.77 mm (irregular), with small RMSD (around 2 mm). Top-of-bed camera position was slightly more accurate, but bottom positions stayed acceptable — important because the ICU staff need free space around the child.
What this means for me
It shows I can go from an idea, to a working system, to protected intellectual property.
It shows I can work with doctors and engineers on one real problem.
And it shows my research has real clinical value, in a real hospital.
This is the kind of applied, useful work I want to teach and lead.
Use ← / → keys, or click the right / left half of the screen.