Butterfly Network‘s Jonathan Rothberg wants to make a “super-low-cost version of a $6 million (ultrasound) machine, to make it 1,000 times cheaper, 1,000 times faster, and a hundred times more precise.” This will depend on software and extensive AI image research to extract key features to automate diagnoses.
Butterfly’s patent applications describe compact, hand held ultrasound scanners that create 3D images in real time. Rothberg want to create a cheap system that can be used in the poorest nations.
Current ultrasound machines use piezoelectric crystals or ceramics to generate and receive sound waves. They are wired and attached with cables to a signal processing box. Butterfly wants to integrate ultrasound elements on a computer chip, cheaply produce them, and simplify the creation of the arrays needed to produce 3D images.