Yang Ling, Peng Shengguang, Yahya Rebaz Othman, Qian Leren
School of Informatics, Harbin Guangsha College, Harbin, 150025, Heilongjiang, China.
School of Engineering and Management, Pingxiang University, Pingxiang, 337055, Jiangxi, China.
J Cancer Res Clin Oncol. 2023 Nov;149(14):13331-13344. doi: 10.1007/s00432-023-05191-2. Epub 2023 Jul 24.
Diagnosis of cancer in breast cells is an important and vital issue in the field of medicine. In this context, the use of advanced methods such as deep complex neural networks and data mining can significantly improve the accuracy and speed of diagnosis. A hybrid approach that can be effective in breast cancer diagnosis is the use of deep complex neural networks and data mining. Due to their powerful nonlinear capabilities in extracting complex features from data, deep neural networks have a very good ability to detect patterns related to cancer. By analyzing millions of data related to breast cells and recognizing common and unusual patterns in them, these networks are able to diagnose cancer with high accuracy. Also, the use of data mining method plays an important role in this process.
Using data mining algorithms and techniques, useful information can be extracted from the available data and the characteristics of healthy and cancerous cells can be separated. This information can be given as input to the deep neural network to achieve more accurate diagnosis. Another method to diagnose breast cancer is the use of thermography, which we use in this research along with data mining and deep learning.
Thermography uses an infrared camera to record the temperature of the target area. This method of breast cancer imaging is less expensive and completely safe compared to other methods. A total of 187 volunteers including 152 healthy people and 35 cancer patients were evaluated. Each person had ten thermographic images, resulting in a total of 1870 thermographic images. Four alternative deep complex neural network models, namely ResNet18, ResNet50, VGG19, and Xception, were used to identify thermal images, including benign and malignant images.
The evaluation results showed that the use of a combined method based on deep complex neural network and data mining in the diagnosis of cancer in breast cells can bring a significant improvement in the accuracy and speed of diagnosis of this important disease.
乳腺细胞癌的诊断是医学领域中一个重要且关键的问题。在这种背景下，使用诸如深度复杂神经网络和数据挖掘等先进方法可以显著提高诊断的准确性和速度。一种在乳腺癌诊断中有效的混合方法是使用深度复杂神经网络和数据挖掘。由于深度神经网络在从数据中提取复杂特征方面具有强大的非线性能力，它们具有很好的检测与癌症相关模式的能力。通过分析数百万与乳腺细胞相关的数据并识别其中的常见和异常模式，这些网络能够高精度地诊断癌症。此外，数据挖掘方法的使用在这个过程中也起着重要作用。
使用数据挖掘算法和技术，可以从可用数据中提取有用信息，并分离出健康细胞和癌细胞的特征。这些信息可以作为深度神经网络的输入，以实现更准确的诊断。另一种诊断乳腺癌的方法是使用热成像，在本研究中我们将其与数据挖掘和深度学习一起使用。
热成像使用红外相机记录目标区域的温度。与其他方法相比，这种乳腺癌成像方法成本更低且完全安全。总共评估了187名志愿者，其中包括152名健康人和35名癌症患者。每个人有十张热成像图像，总共产生了1870张热成像图像。使用四种替代的深度复杂神经网络模型，即ResNet18、ResNet50、VGG19和Xception，来识别热图像，包括良性和恶性图像。
评估结果表明，在乳腺细胞癌诊断中使用基于深度复杂神经网络和数据挖掘的组合方法可以显著提高这种重要疾病诊断的准确性和速度。
