PtosisDiffusion: a training-free workflow for precisely predicting post-operative appearance in blepharoptosis patients based on diffusion models.

IF 4.6 2区生物学 Q2 CELL BIOLOGY

Frontiers in Cell and Developmental Biology Pub Date : 2024-10-30 eCollection Date: 2024-01-01 DOI:10.3389/fcell.2024.1459336

Shenyu Huang, Jiajun Xie, Boyuan Yang, Qi Gao, Juan Ye

{"title":"PtosisDiffusion: a training-free workflow for precisely predicting post-operative appearance in blepharoptosis patients based on diffusion models.","authors":"Shenyu Huang, Jiajun Xie, Boyuan Yang, Qi Gao, Juan Ye","doi":"10.3389/fcell.2024.1459336","DOIUrl":null,"url":null,"abstract":"Purpose: This study aims to develop a diffusion-based workflow to precisely predict postoperative appearance in blepharoptosis patients.Methods: We developed PtosisDiffusion, a training-free workflow that combines face mesh with ControlNet for accurate post-operative predictions, and evaluated it using 39 preoperative photos from blepharoptosis patients. The performance of PtosisDiffusion was compared against three other diffusion-based methods: Conditional Diffusion, Repaint, and Dragon Diffusion.Results: PtosisDiffusion demonstrated superior performance in subjective evaluations, including overall rating, correction, and double eyelid formation. Statistical analyses confirmed that PtosisDiffusion achieved the highest overlap ratio (0.87 <math><mrow><mo>±</mo></mrow> </math> 0.07) and an MPLPD ratio close to 1 (1.01 <math><mrow><mo>±</mo></mrow> </math> 0.10). The model also showed robustness in extreme cases, and ablation studies confirmed the necessity of each model component.Conclusion: PtosisDiffusion generates accurate postoperative appearance predictions for ptosis patients using only preoperative photographs. Among the four models tested, PtosisDiffusion consistently outperformed the others in both subjective and statistical evaluation.","PeriodicalId":12448,"journal":{"name":"Frontiers in Cell and Developmental Biology","volume":"12 ","pages":"1459336"},"PeriodicalIF":4.6000,"publicationDate":"2024-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11557522/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Frontiers in Cell and Developmental Biology","FirstCategoryId":"99","ListUrlMain":"https://doi.org/10.3389/fcell.2024.1459336","RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/1/1 0:00:00","PubModel":"eCollection","JCR":"Q2","JCRName":"CELL BIOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Purpose: This study aims to develop a diffusion-based workflow to precisely predict postoperative appearance in blepharoptosis patients.

Methods: We developed PtosisDiffusion, a training-free workflow that combines face mesh with ControlNet for accurate post-operative predictions, and evaluated it using 39 preoperative photos from blepharoptosis patients. The performance of PtosisDiffusion was compared against three other diffusion-based methods: Conditional Diffusion, Repaint, and Dragon Diffusion.

Results: PtosisDiffusion demonstrated superior performance in subjective evaluations, including overall rating, correction, and double eyelid formation. Statistical analyses confirmed that PtosisDiffusion achieved the highest overlap ratio (0.87 $\pm$ 0.07) and an MPLPD ratio close to 1 (1.01 $\pm$ 0.10). The model also showed robustness in extreme cases, and ablation studies confirmed the necessity of each model component.

Conclusion: PtosisDiffusion generates accurate postoperative appearance predictions for ptosis patients using only preoperative photographs. Among the four models tested, PtosisDiffusion consistently outperformed the others in both subjective and statistical evaluation.

查看原文本刊更多论文

PtosisDiffusion：基于扩散模型精确预测眼睑下垂患者术后外观的免训练工作流程。

目的：本研究旨在开发一种基于扩散的工作流程，以精确预测眼睑下垂患者的术后外观：我们开发了上睑下垂扩散（PtosisDiffusion）--一种无需训练的工作流程，它将面部网格与 ControlNet 相结合，可准确预测术后情况，并使用 39 张眼睑外翻患者的术前照片对其进行了评估。将 PtosisDiffusion 的性能与其他三种基于扩散的方法进行了比较：结果：结果表明：上睑下垂扩散法在主观评价（包括总体评分、校正和双眼皮形成）方面表现出色。统计分析证实，PtosisDiffusion 的重叠率最高（0.87 ± 0.07），MPLPD 比率接近 1（1.01 ± 0.10）。该模型在极端情况下也表现出稳健性，消融研究证实了模型各组成部分的必要性：结论：PtosisDiffusion 仅使用术前照片就能准确预测上睑下垂患者的术后外观。在测试的四个模型中，PtosisDiffusion 在主观和统计评估方面的表现始终优于其他模型。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Frontiers in Cell and Developmental Biology Biochemistry, Genetics and Molecular Biology-Cell Biology

CiteScore

9.70

自引率

3.60%

发文量

2531

审稿时长

12 weeks

期刊介绍： Frontiers in Cell and Developmental Biology is a broad-scope, interdisciplinary open-access journal, focusing on the fundamental processes of life, led by Prof Amanda Fisher and supported by a geographically diverse, high-quality editorial board. The journal welcomes submissions on a wide spectrum of cell and developmental biology, covering intracellular and extracellular dynamics, with sections focusing on signaling, adhesion, migration, cell death and survival and membrane trafficking. Additionally, the journal offers sections dedicated to the cutting edge of fundamental and translational research in molecular medicine and stem cell biology. With a collaborative, rigorous and transparent peer-review, the journal produces the highest scientific quality in both fundamental and applied research, and advanced article level metrics measure the real-time impact and influence of each publication.