Skip to main content
Usami Laboratory
Research explainers

Anatomy-Preserving Augmentation via Generative AI

A workflow for augmenting medical training images with anatomical controls and generated-sample selection.

Proceedings of the 19th International Joint Conference on Biomedical Engineering Systems and Technologies -- BIOIMAGING · 2026

  1. 01

    Prompt Generation

  2. 02

    Anatomy-Preserving Synthesis

  3. 03

    Segmentation Evaluation

Conceptual overview of the research workflow

The problem

Simple transformations risk distorting anatomical structures when addressing the scarcity of medical image learning data.

The research idea

A reproducible workflow was constructed using Stable Diffusion and ControlNet, covering prompt generation to evaluation.

What was evaluated

Quantitative evaluation using IoU and Dice coefficients was conducted on retinal DRIVE and carotid ultrasound IMT segmentation.

Considering applications

Considered as a candidate for mitigating training data scarcity in medical image analysis models.

Original research

Anatomy-Preserving Diffusion-Based Data Augmentation for Medical Image Segmentation

Tominari, Hiroki, Nakayama, Natsuki, Arakawa, Naoko, Usami, Hiroyasu

Proceedings of the 19th International Joint Conference on Biomedical Engineering Systems and Technologies -- BIOIMAGING, 2026 · BIOIMAGING 2026

Publication recordPublic sourcePDF