Introduce a known reference.
A repeatable physical object gives the machine something precise to reveal.
Physics measured. Machine specific.
See what the
machine could see.
Smearlets maps how an imaging system blurs and stretches reality—then uses those measurements to recover sharper, more trustworthy output.
Not AI-generated. No learned features. No hallucinated anatomy.
Distortion field / live studyMove cursor to correct
MEASURED
CORRECTION
CORRECTION
01 / The problem
Blur does not
announce itself.
Every imaging machine has limitations that distort reality. Some information is softened, stretched, or hidden before a clinician ever sees the image.
2M
estimated new cancer diagnoses in the United States in 2025—about 5,600 every day.
Early detection and precise treatment depend on image quality. Resolution can be the difference between detection at an early stage and detection at an advanced stage.
02 / The method
Give distortion
a coordinate system.
We place a known calibration object in the imaging field. How that object changes tells us what the machine is doing at every point.
Calibration widget300 metal spheres
known geometry
local blur
orientation
Measure the deformation.
Each sphere becomes a local measurement of scale, anisotropy, and orientation.
Build the correction field.
Mathematics converts those measurements into a machine-specific map.
Recover hidden resolution.
The correction layer improves output using measured physics, not generated features.
03 / Proof of concept
A sharper image,
without inventing one.
Demonstrated on linear accelerator (LINAC) images. Drag the divider to inspect the original and corrected output.
↔
ORIGINAL
DECONVOLVED
7×–72×targeted correction
0training datasets required
1machine-specific physical model
04 / Product vision
One principle.
Many machines.
Smearlets is envisioned as a calibration kit and software platform. The widget adapts to the machine; the underlying approach remains consistent.
SMEARLETS
correction layer
from first beachhead to broader platform reach
First reachable LINAC installations,
cancer centers, research hospitals
cancer centers, research hospitals
X-ray oncology imaging
+ radiotherapy imaging
+ radiotherapy imaging
Global medical imaging
systems + software
systems + software
The near-term reachable market expands into broader oncology imaging, then into the larger medical imaging systems and software opportunity.
Can generalize or hallucinate features.
Our approach is anchored to a measured physical reference.
Assumes the image behaves uniformly.
We map local distortion across the field.
Can be inherently expensive.
We are building an adaptable calibration and software layer.
05 / Team
Built at the intersection of
physics and software.
CEO / Medical Physics
Zachary Mullaghy
Medical Physics M.Sc. focused on novel mathematical methods for medical imaging.
COO / Astrophysics + Software
Samantha Goldwasser
Astrophysics M.Sc. specializing in spectral analysis methods and software development.NEXT
06 / What’s next
From one machine to many.
We are expanding Smearlets from a successful LINAC demonstration toward multi-machine validation and a broader medical-imaging platform.
We are open to conversations with investors, research partners, medical institutions, and strategic collaborators who see where this can go.