Scholarlink

This paper proposes a new methodology which fuses Convolutional Neural Networks (CNNs) to embed spatial information and Recurrent Neural Network (RNNs) to incorporate temporal context, achieving the detection of osteoporosis. Our model improves prediction of osteoporosis by combining spatial features from bone images with temporal progression data. Moreover, patient covariates including age, sex, and medical history have been integrated for the sake of model robustness. This method outperforms traditional approaches; the new proposed method provides a comprehensive approach to aid in early diagnosis and individualized planning for osteoporosis treatment. This method improves the diagnostic sensitivity for osteoporosis. The proposed approach has achieved a 7% better accuracy than traditional CNN models.

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

A Hybrid CNN-RNN Approach for Accurate Prediction of Osteoporosis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • R. Geetha,
  • S. Arulselvi,
  • R. Tamilselvi,
  • M. Parisa Beham,
  • S. Nandhineeswari

摘要

This paper proposes a new methodology which fuses Convolutional Neural Networks (CNNs) to embed spatial information and Recurrent Neural Network (RNNs) to incorporate temporal context, achieving the detection of osteoporosis. Our model improves prediction of osteoporosis by combining spatial features from bone images with temporal progression data. Moreover, patient covariates including age, sex, and medical history have been integrated for the sake of model robustness. This method outperforms traditional approaches; the new proposed method provides a comprehensive approach to aid in early diagnosis and individualized planning for osteoporosis treatment. This method improves the diagnostic sensitivity for osteoporosis. The proposed approach has achieved a 7% better accuracy than traditional CNN models.