基于改进神经网络的肺炎图像分类研究

Research on Pneumonia Image Classification Based on Improved Neural Network

摘要: 早期的发现和诊断对于治疗新冠患者至关重要，利用卷积神经网络识别肺部X光图像判别新冠患病与否，在实际医疗中得到了广泛应用。基于卷积神经网络的方法能够迅速、准确地判别肺部X光图像。然而，传统的卷积神经网络模型在处理图像数据时存在一定的不足，特别是在特征提取方面缺乏针对性。为此，本文提出了一种融合金字塔池化模型(PPM)的神经网络模型。本文将DenseNet121模型与PPM特征提取模块进行了巧妙的融合，并通过在肺炎公开数据集上进行验证，展示了该方法在实际应用中的有效性。实验结果表明，本文提出的融合金字塔池化模型的网络架构显著提升了对新冠肺炎的识别准确性。这一创新性的图像识别方法不仅在实践中取得了显著的效果，而且为深度学习在医学影像领域的应用提供了有益的参考。这对于改善COVID-19早期诊断和治疗具有积极的推动作用，对于未来类似疾病的防控和医学研究也具有一定的指导意义。

Abstract: Early detection and diagnosis are crucial for the treatment of COVID-19 patients. The use of convo-lutional neural networks to identify lung X-ray images and determine whether a person is infected with COVID-19 has been widely applied in practical medical settings. Methods based on convolu-tional neural networks can rapidly and accurately discern lung X-ray images. However, traditional convolutional neural network models have certain limitations when processing image data, partic-ularly in the lack of specificity in feature extraction. To address this issue, this paper proposes a neural network model that integrates a Pyramid Pooling Module (PPM). The paper ingeniously combines the DenseNet121 model with the PPM feature extraction module. Through validation on a public dataset of pneumonia cases, the effectiveness of this method in practical applications is demonstrated. Experimental results show that the network architecture of the proposed integrated Pyramid Pooling Model significantly improves the accuracy of identifying COVID- 19 pneumonia. This innovative image recognition method not only achieves remarkable results in practice but also provides valuable references for the application of deep learning in the field of medical imaging. It plays a positive role in advancing the early diagnosis and treatment of COVID- 19, and it has signifi-cant implications for the prevention and medical research of similar diseases in the future.