Can LLM-Generated Text Empower Surgical Vision-Language Pre-training?
Paper • 2604.18134 • Published
Access the Dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
Terms of Access
You need to agree to share your contact information to access this dataset.
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
⚠️ Important Requirement:
To verify your research purpose, you must use a HuggingFace account registered with an official email address associated with your academic research institution to request access.
Requests using personal email domains (e.g., Gmail, Yahoo, Outlook) will be rejected.
Conditions:
- I agree that using the Database only for non-commercial research and educational purposes.
- I agree that King's College London make no representations or warranties regarding the Database, including but not limited to warranties of non-infringement or fitness for a particular purpose.
- I agree that full responsibility for his or her use of the Database and shall defend and indemnify the LIME team and King's College London, including their employees, Trustees, officers, and agents, against any and all claims arising from the Researcher's use of the Database, including but not limited to the Researcher's use of any copies of copyrighted images that he or she may create from the Database.
- I agree that I may provide research associates and colleagues with access to the Database provided that they first agree to be bound by these terms and conditions.
- I agree that King's College London reserve the right to terminate the Researcher's access to the Database at any time.
- I agree that if the Researcher is employed by a for-profit, commercial entity, the Researcher's employer shall also be bound by these terms and conditions, and the Researcher hereby represents that he or she is fully authorized to enter into this agreement on behalf of such employer.
- I agree that the law of the United Kingdom of Great Britain and Northern Ireland shall apply to all disputes under this agreement.
Log in or Sign Up to review the conditions and access this dataset content.
Can LLM-Generated Text Empower Surgical Vision-Language Pre-training? (CVPRW 2026 Oral)
This is the official repository for the CVPRW 2026 Oral paper: Can LLM-Generated Text Empower Surgical Vision-Language Pre-training?
Star ⭐ the repository if you find this work useful.
Abstract
Recent advancements in self-supervised learning have led to powerful surgical vision encoders capable of spatiotemporal understanding. However, extending these visual foundations to multimodal reasoning tasks is severely bottlenecked by the prohibitive cost of expert textual annotations.
To overcome this scalability limitation, we introduce LIME, a large-scale multimodal dataset derived from open-access surgical videos using human-free, Large Language Model (LLM)-generated narratives. While LIME offers immense scalability, unverified generated texts may contain errors, including hallucinations, that could potentially lead to catastrophically degraded pretrained medical priors in standard contrastive pipelines.
To mitigate this issue, we propose SurgLIME, a parameter-efficient Vision-Language Pre-training (VLP) framework designed to learn reliable cross-modal alignments from noisy narratives. SurgLIME preserves foundational medical priors using a LoRA-adapted dual-encoder architecture and introduces an automated confidence estimation mechanism that dynamically down-weights uncertain text during contrastive alignment.
Evaluations on the AutoLaparo and Cholec80 benchmarks show that SurgLIME achieves competitive zero-shot cross-modal alignment while preserving the robust linear-probing performance of the visual foundation model.
Installation
Clone the repository and install the required dependencies:
git clone git@github.com:visurg-ai/SurgLIME.git
cd SurgLIME
pip install -r requirements.txt
- Downloads last month
- 20