← Back to all projects

DeepNerve-3D

A private research prototype for mandibular nerve segmentation in 3D CBCT images.

Sources reviewed: 2026-09-12

01 / CONTEXT

The problem

Tracing a thin anatomical structure in a 3D scan requires attention to continuity as well as local appearance.

02 / OWNERSHIP

My contribution

  • Led the project requirements and choices, with AI-assisted implementation and review.
Working approach: product-led, AI-assisted implementation, followed by review and iteration.
03 / ENGINEERING JUDGMENT

Decisions & tradeoffs

01

Keep the public scope narrow

The choice
Describe only the concept already published on the portfolio.
The tradeoff
The implementation cannot be assessed from this page alone.
How it is checked
The linked earlier portfolio snapshot records the public description.

What the work shows

The public overview names SwinUNETR and topology-aware post-processing as the project direction; model results are not published here.

FROM CLAIM TO SOURCE

Follow the evidence

What this does not establish

  • Source code is not public. Only the existing public description is used.
  • No private data, unpublished results, or independently verified deployment claim is included.

Tools & methods

3D imagingPrivate project