Scholarlink

This paper presents a comprehensive prototype of a tool for generating and validating patches and labels in YOLO (You Only Look Once) format from whole-slide images (WSIs) of the Pap test. The system combines patch and label generation in Qupath with a Groovy script and visual validation of labels using OpenCV. One of the most innovative contributions is the development of a web application, utilizing Flask, for expert review of the validated patches generated in Qupath. This approach generates a method for quality control of dataset development for training Deep Learning models from Pap test WSI annotations. While still in the prototype phase, this tool demonstrates significant potential as a foundation for developing robust and scalable clinical applications of artificial intelligence in digital cervical cytology.

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

Prototype of a Comprehensive System for Automated Generation and Expert Validation of Labeled Patches on Papanicolaou Test WSI Images for Semi-supervised Training of YOLO Models in Automated Cervical Cytology Diagnosis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Cardona-Mendoza Andrés,
  • Gil-González Ana-Belén,
  • Hector Javier Hortua,
  • Rossi Labianca Lautaro,
  • Perdomo-Lara Sandra

摘要

This paper presents a comprehensive prototype of a tool for generating and validating patches and labels in YOLO (You Only Look Once) format from whole-slide images (WSIs) of the Pap test. The system combines patch and label generation in Qupath with a Groovy script and visual validation of labels using OpenCV. One of the most innovative contributions is the development of a web application, utilizing Flask, for expert review of the validated patches generated in Qupath. This approach generates a method for quality control of dataset development for training Deep Learning models from Pap test WSI annotations. While still in the prototype phase, this tool demonstrates significant potential as a foundation for developing robust and scalable clinical applications of artificial intelligence in digital cervical cytology.