Scholarlink

<p>Polycystic ovary syndrome (PCOS) is an ever-prevalent and ever-challenging endocrine disorder that affects millions of women, leading to infertility and metabolic complications, oftentimes after delayed diagnosis from inconsistent clinical presentations. The traditional diagnostic methods, such as the Rotterdam Criteria, rely on subjective assessment techniques, often resulting in difficulties in early detection. This study introduces an AI-based polycystic ovary syndrome (PCOS) diagnosis and medical report generation mechanism that integrates You Only Look Once version 11 (YOLO v11) with hybrid optimization of Satin Bowerbird Optimization (SBO) and Aquila Optimization (AO) for ovarian cyst detection, and Pali-Gemma v2 for medical imaging interpretation and report generation. Pali-Gemma v2, a vision-language AI model capable of generating structured medical reports with classification of PCOS severity types and other associated ovarian abnormalities. Various pre-processing approaches, such as noise filtering, contrast enhancement, edge detection, and segmentation filters, were applied to the ultrasound images to increase performance. This framework considers an advanced metric evaluation like precision, recall, F1-score for detection, and BLEU, ROUGE, and expert medical for report generation. The proposed system presents a practical application involving an AI-based tool for PCOS detection that would contribute to reducing the delay in diagnostics while ensuring greater efficiency in clinical workflow and providing accurate and interpretable medical insights.</p>

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

Automated PCOS Diagnosis and Medical Report Generation Using YOLO v11 and Pali-Gemma v2

  • 期刊论文

  • 发表日期：

    2025年6月23日

查看全文
  • Neha Yadav,
  • Sagar Dhanraj Pande,
  • A. Ranjith Kumar

摘要

Polycystic ovary syndrome (PCOS) is an ever-prevalent and ever-challenging endocrine disorder that affects millions of women, leading to infertility and metabolic complications, oftentimes after delayed diagnosis from inconsistent clinical presentations. The traditional diagnostic methods, such as the Rotterdam Criteria, rely on subjective assessment techniques, often resulting in difficulties in early detection. This study introduces an AI-based polycystic ovary syndrome (PCOS) diagnosis and medical report generation mechanism that integrates You Only Look Once version 11 (YOLO v11) with hybrid optimization of Satin Bowerbird Optimization (SBO) and Aquila Optimization (AO) for ovarian cyst detection, and Pali-Gemma v2 for medical imaging interpretation and report generation. Pali-Gemma v2, a vision-language AI model capable of generating structured medical reports with classification of PCOS severity types and other associated ovarian abnormalities. Various pre-processing approaches, such as noise filtering, contrast enhancement, edge detection, and segmentation filters, were applied to the ultrasound images to increase performance. This framework considers an advanced metric evaluation like precision, recall, F1-score for detection, and BLEU, ROUGE, and expert medical for report generation. The proposed system presents a practical application involving an AI-based tool for PCOS detection that would contribute to reducing the delay in diagnostics while ensuring greater efficiency in clinical workflow and providing accurate and interpretable medical insights.