흉부 X선 영상을 이용한 작은 층수 ResNet 기반 폐렴 진단 모델의 성능 평가
Performance Evaluation of ResNet-based Pneumonia Detection Model with the Small Number of Layers Using Chest X-ray Images
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
In this study, pneumonia identification networks with the small number of layers were constructed by using chest X-ray images. The networks had similar trainable-parameters, and the performance of the trained models was quantitatively evaluated with the modification of the network architectures. A total of 6 networks were constructed: convolutional neural network (CNN), VGGNet, GoogleNet, residual network with identity blocks, ResNet with bottleneck blocks and ResNet with identity and bottleneck blocks. Trainable parameters for the 6 networks were set in a range of 273,921-294,817 by adjusting the output channels of convolution layers. The network training was implemented with binary cross entropy (BCE) loss function, sigmoid activation function, adaptive moment estimation (Adam) optimizer and 100 epochs. The performance of the trained models was evaluated in terms of training time, accuracy, precision, recall, specificity and F1-score. The results showed that the trained models with the small number of layers precisely detect pneumonia from chest X-ray images. In particular, the overall quantitative performance of the trained models based on the ResNets was above 0.9, and the performance levels were similar or superior to those based on the CNN, VGGNet and GoogleNet. Also, the residual blocks affected the performance of the trained models based on the ResNets. Therefore, in this study, we demonstrated that the object detection networks with the small number of layers are suitable for detecting pneumonia using chest X-ray images. And, the trained models based on the ResNets can be optimized by applying appropriate residual-blocks.
Abstract
In this study, pneumonia identification networks with the small number of layers were constructed by using chest X-ray images. The networks had similar trainable-parameters, and the performance of the trained models was quantitatively evaluated with the modification of the network architectures. A total of 6 networks were constructed: convolutional neural network (CNN), VGGNet, GoogleNet, residual network with identity blocks, ResNet with bottleneck blocks and ResNet with identity and bottleneck blocks. Trainable parameters for the 6 networks were set in a range of 273,921-294,817 by adjusting the output channels of convolution layers. The network training was implemented with binary cross entropy (BCE) loss function, sigmoid activation function, adaptive moment estimation (Adam) optimizer and 100 epochs. The performance of the trained models was evaluated in terms of training time, accuracy, precision, recall, specificity and F1-score. The results showed that the trained models with the small number of layers precisely detect pneumonia from chest X-ray images. In particular, the overall quantitative performance of the trained models based on the ResNets was above 0.9, and the performance levels were similar or superior to those based on the CNN, VGGNet and GoogleNet. Also, the residual blocks affected the performance of the trained models based on the ResNets. Therefore, in this study, we demonstrated that the object detection networks with the small number of layers are suitable for detecting pneumonia using chest X-ray images. And, the trained models based on the ResNets can be optimized by applying appropriate residual-blocks.
참고문헌 (33)
- Quekel LG, "Miss rate of lung cancer on the chest radiograph in clinical practice", Chest, 15(3), pp.720~724 (1999)
- 홍준용, "AI 의료영상 분석의 개요 및 연구 현황에 대한 고찰", 방사선기술과학, 43(3), pp.195~208 (2020), 대한방사선과학회
- 김영재, "의료 영상에 최적화된 딥러닝 모델의 개발", 대한영상의학회지, 81(6), pp.1274~1289 (2020), 대한영상의학회[DOI]
- 오주영, "게이트심장혈액풀검사에서 딥러닝 기반 좌심실 영역 분할방법의 유용성 평가", 방사선기술과학, 45(2), pp.151~158 (2022), 대한방사선과학회
- Seah JCY, "Effect of a comprehensive deep-learning model on the accuracy of chest x-ray interpretation by radiologists : A retrospective, multireader multicase study", The Lancet Digital Health, 3(8), pp.496~506 (2021)
- Zakaria R, "Deep learning in medical imaging: A review", Applications of Machine Intelligence in Engineering, pp.131~144 (2022)
- Anis S, "An overview of deep learning approaches in chest radiograph", IEEE Access, 8, pp.182347~182354 (2020)
- Aslani S, "Utilisation of deep learning for COVID-19 diagnosis", Clinical Radiology, 78(2), pp.150~157 (2023)
- Hussain E, "CoroDet : A deep learning based classification for COVID-19 detection using chest X-ray images", Chaos, Solitons & Fractals, 142, pp.110495~ (2021)
- Ikechukwu AV, "ResNet-50 vs VGG-19 vs training from scratch : A comparative analysis of the segmentation and classification of Pneumonia from chest X-ray images", Global Transitions Proceedings, 2(2), pp.375~381 (2021)
- He K, "Deep residual learning for image recognition", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.6~770 (2016)
- Simonyan K, "Very deep convolutional networks for large-scale image recognition", arXiv:1409.1556, ~ (2014)
- Ying X, "An overview of overfitting and its solutions", Journal of Physics : Conference Series, 1168(2), pp.022022~ (2019)
- Yuan P, "HS-ResNet: Hierarchical-split block on convolutional neural network", arXiv:2010.07621, ~ (2020)
- Amjoud AB, "Convolutional neural networks backbones for object detection", Proceedings of the International Conference on Image and Signal Processing, pp.282~289 (2020)
- He T, "Bag of tricks for image classification with convolutional neural networks", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.558~567 (2019)
- Taylor L, "Improving deep learning with generic data augmentation", IEEE Symposium Series on Computational Intelligence, pp.1542~1547 (2018)
- "Pneumonia classification"[URL]
- Shorten C, "A survey on image data augmentation for deep learning", Journal of Big Data, 6(1), pp.1~48 (2019)
- Zagoruyko S, "Wide residual networks", rXiv: 1605.0714, ~ (2016)
- Gu J, "Recent advances in convolutional neural networks", Pattern Recognition, 77, pp.354~377 (2018)
- Szegedy C, "Going deeper with convolutions", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.1~9 (2015)
- Loshchilov I, "Decoupled weight decay regularization", arXiv: 1711.05101, ~ (2017)
- Dhumale RB, "Machine learning: A way of dealing with Artificial Intelligence", International Conference on Innovations in Information and Communication Technology, pp.1~6 (2019)
- Gholamalinezhad H, "Pooling methods in deep neural networks, a review", arXiv: 2009. 07485, ~ (2020)
- Jiao Y, "Performance measures in evaluating machine learning based bioinformatics predictors for classifications", Quantitative Biology, 4, pp.320~330 (2016)
- Lin TY, "Focal loss for dense object detection", Proceedings of the IEEE International Conference on Computer Vision, pp.2980~2988 (2017)
- Nagi J, "Max-pooling convolutional neural networks for vision-based hand gesture recognition", IEEE International Conference on Signal and Image Processing Applications, pp.342~347 (2011)
- Guler O, "Classification performance of deep transfer learning methods for pneumonia detection from chest X-ray images", Journal of Artificial Intelligence and Systems, 4(1), pp.107~126 (2022)
- Chouhan V, "A novel transfer learning based approach for pneumonia detection in chest X-ray images", Applied Sciences, 10(2), pp.559~ (2020)
- Ali F, "A two-tier framework based on GoogLeNet and YOLOv3 Models for tumor detection in MRI", Computers, Materials and Continua, 72, pp.73~92 (2022)
- Han Y, "Pneumonia detection on chest x-ray using radiomic features and contrastive learning", IEEE International Symposium on Biomedical Imaging, pp.247~251 (2021)
- Yu Z, "Research on face recognition classification based on improved GoogleNet", Security and Communication Networks, 2022, pp.1~6 (2022)