Scholarlink

<p>Proper, precise and early diagnosis of lung diseases is very important for effective treatment and patient outcomes. In this research article, we present a novel approach for lung malignancy prediction utilizing Short-Time Fourier Transform (STFT) for feature extraction and Bidirectional Long Short-Term Memory (Bi-LSTM) networks for classification. The proposed methodology leverages the spectral analysis capabilities of STFT to transform lungs image into time–frequency representations, capturing critical patterns associated with various lung malignancy conditions. These features are then fed into a Bi-LSTM network, which excels in modeling temporal dependencies in sequential data, providing a robust framework for predicting lung malignancy. Extensive experiments were conducted on a dataset comprising normal and pathological cases. The findings of the proposed method demonstrate that the given approach outperforms traditional machine learning models and achieves high accuracy in distinguishing between healthy and diseased states. The Bi-LSTM network's ability to learn bidirectional temporal features contributes to improved sensitivity and specificity in detecting subtle anomalies in lungs. This study highlights the potential of combining STFT and Bi-LSTM for developing efficient, non-invasive diagnostic tools, bringing a ray of hope for advancements in healthcare and telemedicine applications.</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

Predicting lung cancer using short-time fourier transform and bidirectional long short-term memory networks

  • 期刊论文

  • 发表日期：

    2025年5月21日

查看全文
  • Sneha Khoria,
  • Raghuraj Singh

摘要

Proper, precise and early diagnosis of lung diseases is very important for effective treatment and patient outcomes. In this research article, we present a novel approach for lung malignancy prediction utilizing Short-Time Fourier Transform (STFT) for feature extraction and Bidirectional Long Short-Term Memory (Bi-LSTM) networks for classification. The proposed methodology leverages the spectral analysis capabilities of STFT to transform lungs image into time–frequency representations, capturing critical patterns associated with various lung malignancy conditions. These features are then fed into a Bi-LSTM network, which excels in modeling temporal dependencies in sequential data, providing a robust framework for predicting lung malignancy. Extensive experiments were conducted on a dataset comprising normal and pathological cases. The findings of the proposed method demonstrate that the given approach outperforms traditional machine learning models and achieves high accuracy in distinguishing between healthy and diseased states. The Bi-LSTM network's ability to learn bidirectional temporal features contributes to improved sensitivity and specificity in detecting subtle anomalies in lungs. This study highlights the potential of combining STFT and Bi-LSTM for developing efficient, non-invasive diagnostic tools, bringing a ray of hope for advancements in healthcare and telemedicine applications.