Mission Balance: Generating Under-Represented Class Samples Using Video Diffusion Models

doi:10.1007/978-3-032-05141-7_40

2026-01-01

PRE

Danush Kumar Venkatesh *

Isabel Funke

Micha Pfeiffer

Fiona R. Kolbinger

Hanna Maria Schmeiser

Marius Distler

Jürgen Weitz

Stefanie Speidel

DOI：10.1007/978-3-032-05141-7_40

原文链接

原文求助

摘要

En 中文

计算机辅助干预可通过深度学习方法利用手术视频中的时空信息来改进术中指导。然而，手术视频数据集中常见的数据严重不平衡问题阻碍了高性能模型的开发。本研究旨在通过合成手术视频来克服数据不平衡问题。我们提出了一种独特的两阶段、文本条件扩散方法，用于为代表性不足的类别生成高保真手术视频。我们的方法将生成过程基于文本提示，并通过利用二维潜在扩散模型捕捉空间内容，然后集成时序注意力层来确保时序一致性，从而解耦空间和时序建模。此外，我们引入了一种拒绝采样策略来选择最合适的合成样本，有效扩充现有数据集以解决类别不平衡问题。我们在两个下游任务——手术动作识别和术中事件预测——上评估了我们的方法，结果表明，采用我们方法生成的合成视频能显著提升模型性能。

Keyword:

Video Diffusion

Surgical Data Science

Data Imbalance

期刊

MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2025, PT XI

IF:

论文数:

被引数:

机构

helmholtz association

学者数:

6.3K

论文数: 2.3K

被引数: 6

german cancer research center (dkfz)

学者数:

313

论文数: 149

被引数: 0

引用论文

Rethinking FID: Towards a Better Evaluation Metric for Image Generation 重新思考FID: 为图像生成提供更好的评估指标

2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

IF0

2024-06-16

PREAI

Sadeep Jayasumana; Srikumar Ramalingam; Andreas Veit; Daniel Glasner; Ayan Chakrabarti; Sanjiv Kumar

U-Net: Convolutional Networks for Biomedical Image Segmentation U-net: 用于生物医学图像分割的卷积网络

Lecture Notes in Computer Science

IF0

2015-11-18

PREAI

Olaf Ronneberger; Philipp Fischer; Thomas Brox

X3D: Expanding Architectures for Efficient Video Recognition X3D: 扩展高效视频识别架构

2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

IF0

2020-06-01

OAAI

Christoph Feichtenhofer

Long Short-Term Memory 长短期记忆

Neural Computation

IF0

1997-11-01

PREAI

Sepp Hochreiter; Jürgen Schmidhuber

High-Resolution Image Synthesis with Latent Diffusion Models 具有潜在扩散模型的高分辨率图像合成

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

IF0

2022-06-01

OAAI

Robin Rombach; Andreas Blattmann; Dominik Lorenz; Patrick Esser; Bjorn Ommer

Adding Conditional Control to Text-to-Image Diffusion Models 将条件控制添加到文本到图像扩散模型

2023 IEEE/CVF International Conference on Computer Vision (ICCV)

IF0

2023-10-01

OAAI

Lvmin Zhang; Anyi Rao; Maneesh Agrawala

Surgical data science-from concepts toward clinical translation 外科数据科学-从概念到临床翻译

MEDICAL IMAGE ANALYSIS

IF11.8

2022-02-01

135

OAAI

Maier-Hein, Lena; Eisenmann, Matthias; Sarikaya, Duygu; Maerz, Keno; Collins, Toby; Malpani, Anand; Fallert, Johannes; Feussner, Hubertus; Giannarou, Stamatia; Mascagni, Pietro; Nakawala, Hirenkumar; Park, Adrian; Pugh, Carla; Stoyanov, Danail; Vedula, Swaroop S.; Cleary, Kevin; Fichtinger, Gabor; Forestier, Germain; Gibaud, Bernard; Grantcharov, Teodor; Hashizume, Makoto; Heckmann-Noetzel, Doreen; Kenngott, Hannes G.; Kikinis, Ron; Muendermann, Lars; Navab, Nassir; Onogur, Sinan; Ross, Tobias; Sznitman, Raphael; Taylor, Russell H.; Tizabi, Minu D.; Wagner, Martin; Hager, Gregory D.; Neumuth, Thomas; Padoy, Nicolas; Collins, Justin; Gockel, Ines; Goedeke, Jan; Hashimoto, Daniel A.; Joyeux, Luc; Lam, Kyle; Leff, Daniel R.; Madani, Amin; Marcus, Hani J.; Meireles, Ozanan; Seitel, Alexander; Teber, Dogu; Ueckert, Frank; Mueller-Stich, Beat P.; Jannin, Pierre; Speidel, Stefanie

Synthesising Rare Cataract Surgery Samples with Guided Diffusion Models

Lecture Notes in Computer Science

IF0

2023-10-01

PREAI

Yannik Frisch; Moritz Fuchs; Antoine Sanner; Felix Anton Ucar; Marius Frenzel; Joana Wasielica-Poslednik; Adrian Gericke; Felix Mathias Wagner; Thomas Dratsch; Anirban Mukhopadhyay

Investigation of intraoperative factors associated with postoperative pancreatic fistula following laparoscopic left pancreatectomy with stapled closure: a video review-based analysis 研究腹腔镜左胰腺切除术使用切割闭合器闭合后发生术后胰瘘相关术中因素的调查：一项基于视频回顾的分析

Surgical Endoscopy

IF0

2020-09-10

PREAI

Giuseppe Zimmitti; Roberta La Mendola; Alberto Manzoni; Valentina Sega; Valentina Malerba; Elio Treppiedi; Claudio Codignola; Lorenzo Monfardini; Marco Garatti; Edoardo Rosso

IF0

PREAI

查看更多内容