3D Microscopy: Combining Meta-Optics with Machine Learning

-

During my elementary school years, my biology teacher brought a microscope. We didn’t do a lot of hands-on learning but this was one rare exception. As a curious kid, I jumped at the opportunity to look at it. It was so fascinating to see the tiny universe it had unfolded. And yet as I looked at it, I couldn’t help but feel frustrated that I couldn’t see more; a 2D flat image didn’t do it justice. Little did I know then that traditional microscopes indeed have their problems, and what felt to me as a limitation was indeed something scientists struggled with.

Jump forward to today, humanity still hasn’t found a way to monitor a cell in real-time 3D. Traditional microscopes need you to analyze cell samples on a flat glass slide. Of course, this is ineffective if we want to analyze the real-time behavior of a cell. There have been attempts to create a 3D microscope but aren’t really successful or just inefficient to use.

National Cancer Institute - Osteosarcoma - Bone Cancer Cell
Image: National Cancer Institute / Unsplash

Current 3D microscopy records a sample pixel by pixel to generate a 3D view. This takes a long time and can only take a few shots in a minute which is not enough to keep up with the dynamicity of a cell. It contributed to learning a lot about the microscopic world inside us, but it didn’t quite reach our goal. However, a team of researchers from Tampere University in Finland proposed a way to break this barrier using meta-optics and machine learning.

Capturing cells

By combining meta-optics and machine learning, the researchers hope to create a new method for high-speed volumetric microscopy that can capture the dynamics of living cells like never before. Meta-optics is a relatively new science field that uses flat optical elements, such as small lenses, that are designed to bend light in specific ways while being precise and efficient.

These structures can manipulate light waves at the sub-wavelength level, which allows them to achieve high-resolution imaging of living cells. Unlike in the previous 3D microscopy tools, here a machine learning algorithm will analyze the data, optimize the parameters and remove the outside factors such as noise and distortions to produce a high-quality result.

The “High-speed 3D microscopy” project was slated to commence in April 2023 and extend until the conclusion of 2025, supported by the joint Future Makers Programme of Jane and Aatos Erkko Foundation, along with the Technology Industries of Finland Centennial Foundation, it stands to harness cutting-edge advancements in technology and imaging. Involved in this project are Humeyra Caglayan, Atanas Gotchev, Erdem Sahin, and Teemu Ihalainen.

The implications of such a technology are enormous. Monitoring cell behavior in real-time can lead to many discoveries, both about the cell structure and diseases. Humeyra Caglayan, lead of the project, gives the example of observing cancer cells. Learning about how diseases work exactly is how to combat them. With this and similar innovations, we’re getting close.

Photo credit: The feature image is symbolic and has been done by Christopher Isak with Midjourney for TechAcute. The image in the body of the article has been done by the National Cancer Institute in the US.
Source: Tampere University / Phys.org

Was this post helpful?

Benjamin Adjiovski
Benjamin Adjiovski
Hi! I am a Computer Science Engineer with a passion for all things related to technology. I believe that technology has the power to change the world, so I love staying up-to-date on the latest innovations. If you share the same passion, be my guest.
- Advertisment -
- Advertisment -
- Advertisment -
- Advertisment -
- Advertisment -
- Advertisment -