Scholarlink

This paper details the design and architecture of PARROT (Platform for ARtificial intelligence guided Radiation Oncology Treatment), a free and open-source, Web-based platform that aims to improve therapeutic decision-making through advanced artificial intelligence (AI) techniques. The primary goals of PARROT are to facilitate the visualization of the outputs of AI models, to provide tools for annotating and modifying AI segmentation results, and to assist in comparing treatment plans and selecting the most suitable option. PARROT addresses common challenges in AI research, such as simplifying the execution and management of AI experiments and facilitating the comparison of results between published and institution-trained models. It enables users to interact with AI models for radiation oncology in real time, allowing them to understand and refine model outputs by visually inspecting and correcting biases. PARROT provides an end-to-end workflow with an easy-to-use graphical interface that promotes the implementation of AI models in clinical settings.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

PARROT - A Versatile Platform for AI-Driven Image Segmentation and Dose Prediction

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Wei Zhao,
  • Margerie Huet,
  • Sylvain Deffet,
  • Edmond Sterpin,
  • Ana Barragan Montero,
  • Sébastien Jodogne,
  • John Lee

摘要

This paper details the design and architecture of PARROT (Platform for ARtificial intelligence guided Radiation Oncology Treatment), a free and open-source, Web-based platform that aims to improve therapeutic decision-making through advanced artificial intelligence (AI) techniques. The primary goals of PARROT are to facilitate the visualization of the outputs of AI models, to provide tools for annotating and modifying AI segmentation results, and to assist in comparing treatment plans and selecting the most suitable option. PARROT addresses common challenges in AI research, such as simplifying the execution and management of AI experiments and facilitating the comparison of results between published and institution-trained models. It enables users to interact with AI models for radiation oncology in real time, allowing them to understand and refine model outputs by visually inspecting and correcting biases. PARROT provides an end-to-end workflow with an easy-to-use graphical interface that promotes the implementation of AI models in clinical settings.