Areas of our work
CBCT in dentistry
We are developing CBCT volume analysis workflows, including delineating selected anatomical structures and displaying segmentations over the source image. Our aim is to make results available for review in slices and 3D, and to retain them for further assessment.
A pilot helps evaluate a specific method on agreed material and establish the data preparation and result checking it requires.
Research on spatial images
We prepare experiments to compare methods, computation variants and settings. Our interests include segmentations and maps describing selected properties of volumetric data.
A team can agree on a shared protocol, analyse differences between outputs and revisit the parameters used in individual runs.
Developing new algorithms
Collaboration with a method's author starts by defining its purpose, requirements and limitations. We then prepare its execution workflow, test the data flow and agree on how users should assess the output.
Education
Test images, segmentations and 3D scenes can support learning image analysis and student projects. We agree on the scope with the project supervisor, taking account of available data and educational objectives.
For healthcare organisations
A CBCT analysis pilot tailored to a specific task
Are you considering computer-assisted analysis in your CBCT workflow? Let us start by describing the task: the material you work with, the result you need and how it will be assessed.
We offer pilot collaboration to evaluate a selected method and explore how it could fit into your team's workflow.
What a pilot may include
- Defining the objective and evaluation criteria.
- Checking the technical requirements for the input material.
- Running the agreed method and preparing the result for review.
- Jointly assessing limitations and requirements for further use.
What the team receives
Depending on the agreed scope, the deliverable may be a set of segmentation layers or a scene for use in 3D Slicer, together with information on the method and parameters used. We agree on the delivery format before work begins.
Our current work is developmental and pilot-based. Results require specialist review; we do not offer automated diagnosis or imaging reports through the public website.
How to get started
For an initial conversation, a description of the intended use, data type and expected result is enough. Imaging material is shared only after we agree on access arrangements and how the data will be handled.
For algorithm developers
Prepare your method for a defined workflow
Do you have an image analysis algorithm, a segmentation model or a research method? We invite you to discuss how it can be prepared for repeatable execution and evaluation on agreed data.
We begin with a description of the method: which images it was developed for, what it takes as input, what it returns and its limitations.
Scope of collaboration
Together, we define data formats, dependencies, computing requirements and the output format. We prepare a test protocol and the conditions that each new run of the method must meet.
The code, model and configuration versions should be identifiable so that users know which method produced a result. A version change requires a new assessment of the aspects it affects.
Authors help define the terms of use
Before a method is made available, we agree on rights to the code and model weights, permitted uses, attribution and any commercial terms. Preparing an experiment or a draft version does not automatically publish a service.
A service catalogue in development
Our aim is to let users select methods according to their intended purpose and terms of use, while giving authors a clear route to making their solutions available. Ordering and billing mechanisms are being developed as part of this model.
For now, we offer individually agreed collaboration on integration and pilot use of a method.
For research teams
Experiments with a record of data, methods and results
MedAlgora is developing an environment for running computations on volumetric images and comparing the results. We invite teams working on segmentation, spatial analysis and the evaluation of image processing methods to collaborate.
We start with a research question. Together, we define the input material, analysis variants, metrics and how the results will be reviewed.
What we can prepare together
- An experimental protocol and a description of the parameters.
- A way to run the agreed variants of the computation.
- Links between results, the method version and the input data.
- A results package for further analysis and visualisation.
Comparing methods
Results should be assessed against agreed criteria using comparable material. Within a project, we distinguish differences caused by the method or parameters from those caused by data preparation. We also record cases in which the analysis cannot be completed correctly.
Depending on the task, the package may include segmentations, spatial maps and 3D scenes. The delivery format should let the team inspect the results independently and provide feedback.
Project arrangements
We agree on data access, rights to the results, attribution and the scope of any publication before work begins. Each pilot has a defined scope; subsequent stages depend on its evaluation.
Education and student projects
We invite students and supervisors to work on specific image analysis experiments. The scope is agreed individually.
Explore educational collaboration