Scholarlink

Major Depressive Disorder (MDD) is a prevalent mental illness that requires accurate early diagnosis for effective management. Functional MRI (fMRI) studies have identified altered brain connectivity in regions associated with emotion regulation and mood processing in MDD. We propose an innovative hybrid model for MDD classification from MRI data, combining deep Convolutional Neural Networks (CNNs) and Random Forest. By integrating features from state-of-the-art CNNs VGG16 and EfficientNet, our approach captures complementary discriminative information. A Random Forest classifier then merges these CNN ensemble features, leveraging its capacity to handle high-dimensional and nonlinear data. Evaluated across five MRI datasets, our model outperforms traditional machine learning and single CNN architectures, demonstrating superior accuracy, recall, precision, and F1 scores for Depression classification.

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

CF-Net: A Hybrid CNN-Random Forest Network for Depression Classification in Brain MRI

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Hui Ding,
  • Chong Liu,
  • Yawei Zhang,
  • Bo Li,
  • Yuhang Huang,
  • Jiacheng Lu,
  • Kaiwen Wang,
  • Rongyin Qin,
  • Yuanyuan Shang

摘要

Major Depressive Disorder (MDD) is a prevalent mental illness that requires accurate early diagnosis for effective management. Functional MRI (fMRI) studies have identified altered brain connectivity in regions associated with emotion regulation and mood processing in MDD. We propose an innovative hybrid model for MDD classification from MRI data, combining deep Convolutional Neural Networks (CNNs) and Random Forest. By integrating features from state-of-the-art CNNs VGG16 and EfficientNet, our approach captures complementary discriminative information. A Random Forest classifier then merges these CNN ensemble features, leveraging its capacity to handle high-dimensional and nonlinear data. Evaluated across five MRI datasets, our model outperforms traditional machine learning and single CNN architectures, demonstrating superior accuracy, recall, precision, and F1 scores for Depression classification.