Acoustic imaging has revolutionized medicine over the last few decades; anyone who has had an ultrasound to examine a baby in utero, check an organ or look at a bone has encountered the technology.
It’s not just a medical miracle: The technology is important in mining and resource extraction, manufacturing, building inspection, and dozens of other industries.
While ultrasound in its current form is extremely useful, it does have limits: There’s a big tradeoff between the resolution of the image and the depth of the imaging. To get around this, ultrasound sometimes relies on external “labels,” like special materials or contrast agents deposited near the target area to improve the images. However, it’s often not practical or safe to deposit these labels, meaning there are many situations in which it’s hard to get good ultrasound images.
That’s why Chu Ma, an assistant professor of electrical and computer engineering at the University of Wisconsin-Madison, will use a National Science Foundation CAREER Award to develop a hardware and software acoustic imaging system that can look deeper without relying on these types of labels.
The new system, she says, will allow for imaging that ultrasound cannot currently do, like looking into deep brain tissues, or spotting sub-millimeter kidney stones and lesions on blood vessel walls. It will also enable better sensing to guide underwater vehicles and to help in marine biology, improve quality control inspections in manufacturing, as well as improve imaging in agriculture and mining.
The hardware side of the project will focus on spatial mixing, an image processing technique that uses what are called “blind labels.” These are randomly spaced irregularities, like rocks in the soil or different types of tissues in the human body, that can be used as acoustic scatterers. These blind labels are able to send more signal data back to the sensor.
The second part of the project relies on computational signal processing and image reconstruction algorithms that can use these additional acoustic signals to improve the imaging. The algorithms will use statistical knowledge and data collected by the imaging systems, as well as machine learning and database methods, to decode more information and smooth the signals received by the sensors, creating higher resolution images.
“This is a hardware and software co-design problem,” says Ma. “When we design the software, we will optimize it together with the hardware system so the two can work together nicely.”
After developing the overall framework for this system, Ma and her team will then move into an application-driven validation stage, assessing the individual development needs for biomedical imaging systems, underwater imaging and non-destructive structural inspection. They will then build and test lab-scale versions of each of these imaging systems.
“This research started during my PhD studies,” says Ma. “But then we mainly focused on the hardware part. Now, with this CAREER Award support, we will be able to combine both hardware and software to achieve unprecedented image quality.”
As part of the outreach element of the CAREER Award, Ma is creating three pathways to get younger students involved in interdisciplinary acoustics research. First, she is developing a workshop that she plans to present to high school students to introduce them to the discipline. Second, she is hoping to increase the pipeline of traditionally underrepresented students moving into acoustics research, in particular by using her introduction to acoustics class to encourage students to pursue research in her lab. And third she is creating interactive and hands-on exhibitions for outreach programs open to the general public, such as Wisconsin Science Festival and Engineering EXPO.
Featured photo caption: Graduate student Jinuan Lin (left) and Assistant Professor Chu Ma (right) are working to make acoustic imaging systems that can see deeper without using artificial labels. Credit: Joel Hallberg.