EchoFoundationModel-Survey

This is an updating literature review of the foundation models for echocardiogram. We will include the popular network backbones for video representation learning. Note that we current don't consider the multi-modality scenario such as the vision-language model. Moreover, I list the public datasets that are implemented in the papers.

Echocardiogram

	Paper Title	Year	Institution	Code	Modality	Backbone
1	EchoApex: A General-Purpose Vision Foundation Model for Echocardiography pdf	2024	Siemens Healthineers	-	TTE, TEE, ICE	ViT-base
2	EchoFM: Foundation Model for Generalizable Echocardiogram Analysis pdf	2024	MGH	Github	TTE	ViT-base
3	Echo-Vision-FM: A Pre-training and Fine-tuning Framework for Echocardiogram Video Vision Foundation Model pdf	2024	Northwestern University	-	TTE	ViT-base
4	PanEcho: Complete AI-enabled echocardiography interpretation with multi-task deep learning pdf	2025	Yale University	Github	TTE	2D-CNN
5	EchoFlow: A Foundation Model for Cardiac Ultrasound Image and Video Generation pdf	2025	Imperial College London	Huggingface	TTE	A-VAE

Public datasets

	Database	Modality	Format	Amount	Source
1	MIMIC-IV-ECHO link	TTE	video	525,328	Beth Israel Deaconess Medical Center
2	EchoNet-Dynamic link	TTE(A4C)	video	10,030	Stanford University Hospital
3	EchoNet-LVH link	TTE(PLAX)	video	12,000	Stanford University Hospital
4	CAMUS link	TTE(A2C,A4C)	video	500	University Hospital of St. Etienne
5	TMED link	TTE	video	599	Tufts Medical Center
6	RVENet link	TTE(A4C)	video	3,583	Heart and Vascular Center of Semmelweis University in Budapest