Researchers have developed a new framework using multimodal large language models (MLLMs) to automatically generate bounding box annotations for surgical instruments from existing presence labels. This method aims to overcome the limitations of manual annotation, which is time-consuming and requires expert input. The study evaluated six different MLLMs on laparoscopic cholecystectomy data, demonstrating that a structured multi-stage annotation pipeline achieved a mean Intersection over Union (mIoU) of 0.784, closely approaching human-human agreement. AI
IMPACT This research could significantly speed up the creation of annotated datasets for medical imaging, potentially accelerating AI development in surgical analysis and robotics.
RANK_REASON Research paper published on arXiv detailing a new method for data annotation using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
