Open Access

ARTICLE

Structure-Aware Diffusion Image Outpainting for Echocardiographic Field-of-View Extension

Ruijia He¹, Yixin Hu², Jinze Liu³, Qixin Zhang⁴, Duo Peng⁵, Yanyan Chen^5,*

1 State Key Laboratory of Primate Biomedical Research, Institute of Primate Translational Medicine, Kunming University of Science and Technology, Kunming, China
2 College of Electronics and Information Engineering, Sichuan University, Chengdu, China
3 Central South University of Forestry and Technology, Changsha, China
4 College of Computing and Data Science, Nanyang Technological University, Singapore, Singapore
5 School of Computer Science and Technology, Tongji University, Shanghai, China

* Corresponding Author: Yanyan Chen. Email:

(This article belongs to the Special Issue: Advances in Image Generation: Theories, Architectures, and Applications)

Computers, Materials & Continua 2026, 88(3), 73 https://doi.org/10.32604/cmc.2026.083677

Received 08 April 2026; Accepted 02 June 2026; Issue published 23 July 2026

Abstract

Transthoracic Echocardiography (TTE) often suffers from a limited field of view (FoV), which may obscure peripheral cardiac structures and hinder comprehensive visual assessment. Although FoV extension can be formulated as outpainting, public studies on echocardiographic FoV outpainting remain scarce and are mainly represented by cGAN-based reconstruction methods such as echoGAN. However, in noisy ultrasound images with weak boundaries, the challenge is not only realistic texture synthesis, but also structural continuity across the observed–generated boundary. To address this issue, we propose a structure-aware diffusion framework for echocardiographic FoV outpainting. To the best of our knowledge, this is the first work to introduce diffusion models into this task, moving beyond the existing cGAN-based formulation. Our framework combines a diffusion baseline, a Structural Cue Encoder (SCE) for structure-sensitive conditioning, and Structure-aware Diffusion Learning (SDL) for structural regularization during denoising. Experiments show clear and consistent improvements over cGAN-based reconstruction, producing more coherent structures, smoother transitions, and better FoV extension quality.

---

Keywords

---