Scholarlink

<p>Breast cancer is still a life-threatening and common disease in women globally. Early identification is essential to improve outcomes and lower death rates. Recently, number of innovative approaches has been developed for creating an efficient diagnosis and classification of such a harmful sort of cancer. But there is a lack of accuracy. For improving diagnostic accuracy and minimizing false negatives, this research proposes an optimized deep learning model called LTACNN-ADC-BC (Lightweight Temporal Attention-based Convolutional Neural Network for Automatic Detection with Classification of Breast Cancer). The proposed model analyzes mammography images through Multi-window Savitzky–Golay Filter (MWSGF) for noise removal, followed by feature extraction and classification through a Lightweight Temporal Attention-Based Convolution Neural Network (LTACNN). For further optimization, the model incorporates the Stock Exchange Trading Optimization Algorithm (SETOA) for the fine-tuning of network parameters. The proposed approach is tested with MIAS dataset, real-time dataset, and CBIS-DDSM dataset. Performance indicators reflect remarkable improvements over prior methods, with the model identifying 99.51% accuracy and 0.9992 AUC for MIAS, 99.26% accuracy for the real-time dataset, and 98.54% accuracy with 0.983 AUC for CBIS-DDSM dataset. These performances indicate the efficacy of the LTACNN-ADC-BC method in differentiating between normal, benign, and malignant cases. The LTACNN-ADC-BC model has strong potential for integration into computer-assisted diagnostic systems. It provides improved assistance to radiologists for the diagnosis and screening of breast cancer.</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

Optimized Lightweight Temporal Attention-Based Convolutional Neural Network Espoused Automatic Diagnosis and Classification of Breast Cancer

  • 期刊论文

  • 发表日期：

    2025年12月4日

查看全文
  • K. Sreekala,
  • Jayakrushna Sahoo

摘要

Breast cancer is still a life-threatening and common disease in women globally. Early identification is essential to improve outcomes and lower death rates. Recently, number of innovative approaches has been developed for creating an efficient diagnosis and classification of such a harmful sort of cancer. But there is a lack of accuracy. For improving diagnostic accuracy and minimizing false negatives, this research proposes an optimized deep learning model called LTACNN-ADC-BC (Lightweight Temporal Attention-based Convolutional Neural Network for Automatic Detection with Classification of Breast Cancer). The proposed model analyzes mammography images through Multi-window Savitzky–Golay Filter (MWSGF) for noise removal, followed by feature extraction and classification through a Lightweight Temporal Attention-Based Convolution Neural Network (LTACNN). For further optimization, the model incorporates the Stock Exchange Trading Optimization Algorithm (SETOA) for the fine-tuning of network parameters. The proposed approach is tested with MIAS dataset, real-time dataset, and CBIS-DDSM dataset. Performance indicators reflect remarkable improvements over prior methods, with the model identifying 99.51% accuracy and 0.9992 AUC for MIAS, 99.26% accuracy for the real-time dataset, and 98.54% accuracy with 0.983 AUC for CBIS-DDSM dataset. These performances indicate the efficacy of the LTACNN-ADC-BC method in differentiating between normal, benign, and malignant cases. The LTACNN-ADC-BC model has strong potential for integration into computer-assisted diagnostic systems. It provides improved assistance to radiologists for the diagnosis and screening of breast cancer.