DeepPD obtains better de-aberrated images without requiring optical correction, and more completely estimates wavefronts by accounting for nonlinearities in our deformable mirror and other assumptions not captured in our original analytic approach.

In our new preprint, we combine a deep learning framework with our phase-diverse image acquisition called "DeepPD".

We demonstrated that by acquiring few extra images with added optical aberrations ("phase diversities") wavefront sensing can be achieved using an analytic optimization scheme.

Last year we published a simple experimental method for image-based wavefront sensing in fluorescence microscopy using phase diversity.

opg.optica.org/optica/fullt...

Excited to announce the release of a preprint I have been working in collaboration with Magdalena C. Schneider on to improve phase diversity-based wavefront sensing using deep learning.

arxiv.org/abs/2504.14157

Further, you are shown the fact that teams can accomplish feats - like continuously carrying a 200# sand log for over a mile - that individuals cannot. Building a mindset that is team-oriented will take you farther than being a person who is focused solely on their own priorities.

This order applies even - and especially if you are personally suffering. You are taught to instead look for a teammate to help. If every teammate is looking out for each other instead of themselves, no one is left behind.

These are all capabilities that scientists need to withstand the challenges that accompany the task of discovering new knowledge through an isolating period of continuous failure. If you have never tried BJJ or any martial art - consider this your sign to take the leap and start!

In this article I talk about the role Brazilian Jiu-Jitsu has played in my development as a scientist and scholar. BJJ will humble you no matter how smart and strong you are. It requires resilience and humility. It teaches you to problem solve under pressure.

Optical measurement via the Measure app is by far the most slept-on iphone function

Further, the large axial range means that 3D-TrIm gives an unprecedented view into the motion of single viruses in 3D tissue environments.

This detail originates from the high speed: trajectories are sampled every 1 millisecond, yielding 1000 localizations per second.

3D, two-photon images acquired simultaneously give context to the particle motion enabling new insights into the pre-binding motion and contacts of viral particles.

3D-TrIm decouples acquisition for imaging and particle tracking, allowing both acquisitions to be optimized and enabling multiple orders of magnitude faster tracking that is independent of axial extent. Compare the trajectory detail enabled by 3D-TrIm against conventional spinning disk acquisition:

This is part 2/4 of my Bluesky intro to my work as a microscopist!

You've never seen a virus move like this before: 3D-TrIm (3D Tracking and Imaging) is a revolutionary way to track the motion of viruses and other freely diffusing nanoparticles at high speed over large (>10+ µm) axial ranges

The result is that you can arbitrarily choose the volume rate and interpolate the remaining unscanned voxels which are closely-spaced. This enables you to choose your desired balance between imaging rate and quality.

Read the paper here: opg.optica.org/oe/fulltext....

Raster scanning 1 frame at a time is slow. The more planes in your volume the slower it is - and the larger time difference between the bottom and top planes.

In 3D-FASTR, the raster and focus scanning are performed at the same time at a rate that creates linear tessellating patterns.

Welcome to everyone following me 👋. I think a great way to start off is to share some of my work.

Today I start with 3D-FASTR: A faster alternative to conventional point-scan and random access volumetric imaging approaches.

opg.optica.org/oe/fulltext....