컨볼루션 신경망 모델을 이용한 분류에서 입력 영상의 종류가 정확도에 미치는 영향
The Effect of Type of Input Image on Accuracy in Classification Using Convolutional Neural Network Model
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
The purpose of this study is to classify TIFF images, PNG images, and JPEG images using deep learning, and to compare the accuracy by verifying the classification performance. The TIFF, PNG, and JPEG images converted from chest X-ray DICOM images were applied to five deep neural network models performed in image recognition and classification to compare classification performance. The data consisted of a total of 4,000 X-ray images, which were converted from DICOM images into 16-bit TIFF images and 8-bit PNG and JPEG images. The learning models are CNN models - VGG16, ResNet50, InceptionV3, DenseNet121, and EfficientNetB0. The accuracy of the five convo- lutional neural network models of TIFF images is 99.86%, 99.86%, 99.99%, 100%, and 99.89%. The accuracy of PNG images is 99.88%, 100%, 99.97%, 99.87%, and 100%. The accuracy of JPEG images is 100%, 100%, 99.96%, 99.89%, and 100%. Validation of classification performance using test data showed 100% in accuracy, precision, recall and F1 score. Our classification results show that when DICOM images are converted to TIFF, PNG, and JPEG images and learned through preprocessing, the learning works well in all formats. In medical imaging research using deep learn- ing, the classification performance is not affected by converting DICOM images into any format
Abstract
The purpose of this study is to classify TIFF images, PNG images, and JPEG images using deep learning, and to compare the accuracy by verifying the classification performance. The TIFF, PNG, and JPEG images converted from chest X-ray DICOM images were applied to five deep neural network models performed in image recognition and classification to compare classification performance. The data consisted of a total of 4,000 X-ray images, which were converted from DICOM images into 16-bit TIFF images and 8-bit PNG and JPEG images. The learning models are CNN models - VGG16, ResNet50, InceptionV3, DenseNet121, and EfficientNetB0. The accuracy of the five convo- lutional neural network models of TIFF images is 99.86%, 99.86%, 99.99%, 100%, and 99.89%. The accuracy of PNG images is 99.88%, 100%, 99.97%, 99.87%, and 100%. The accuracy of JPEG images is 100%, 100%, 99.96%, 99.89%, and 100%. Validation of classification performance using test data showed 100% in accuracy, precision, recall and F1 score. Our classification results show that when DICOM images are converted to TIFF, PNG, and JPEG images and learned through preprocessing, the learning works well in all formats. In medical imaging research using deep learn- ing, the classification performance is not affected by converting DICOM images into any format
참고문헌 (23)
- LeCun Y, "Deep learning", Nature, 521(7553), pp.436~444 (2015)
- Krizhevsky A, "Imagenet classification with deep convolutional neural networks", Advances in neural information processing systems, 25, pp.1097~1105 (2012)
- Yap MH, "End-to-end breast ultrasound lesions recognition with a deep learning approach", Medical Imaging 2018:Biomedical Applications in Molecular, Structural, and Functional Imaging, 10578, pp.1057819~ (2018), International Society for Optics and Photonics
- Han S, "A deep learning framework for supporting the classification of breast lesions in ultrasound images", Physics in Medicine & Biology, 62(19), pp.7714~ (2017)
- Sahiner B, "Deep learning in medical imaging and radiation therapy", Medical physics, 46(1), pp.e1~e36 (2019)
- 송하연, "인공지능과 헬스 커뮤니케이션", 언론정보연구, 57(3), pp.196~238 (2020), 언론정보연구소[DOI]
- 양성희, "좌심실비대 진단에서 심장초음파와 심전도검사의 정확성", 한국콘텐츠학회 논문지, 16(2), pp.666~672 (2016), 한국콘텐츠학회[DOI]
- Frohlich, ED, "Left ventricular hypertrophy as a risk factor", Cardiology clinics, 4(1), pp.137~144 (1986)
- "2020", ~
- Alghamdi SS, "Study of cardiomegaly using chest x-ray", Journal of Radiation Research and Applied Sciences, 13(1), pp.460~467 (2020)
- Wang X, "Chestx-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.2097~2106 (2017)
- Tschandl P, "The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions", Scientific data, 5(1), pp.1~9 (2018)
- 권순무, "PACS에서 JPEG2000을 이용한 AAPM CT Performance Phantom영상의 압축에 따른 화질변화", 한국방사선학회논문지, 6(3), pp.217~226 (2012), 한국방사선학회[DOI]
- 고성진, "DICOM 영상과 다양한 형식의 영상 비교", 융합신호처리학회 논문지, 18(2), pp.76~83 (2017), 한국융합신호처리학회[DOI]
- Kabachinski J, "TIFF, GIF, and PNG : get the picture?", Biomedical instrumentation & technology, 41(4), pp.297~300 (2007)
- Simonyan K, "Very deep convolutional networks for large-scale image recognition", arXiv preprint arXiv. 2014;1409.1556, ~ (2014)
- He K, "Deep residual learning for image recognition", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.770~778 (2016)
- Szegedy C, "Rethinking the inception architecture for computer vision", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.2818~2826 (2016)
- Huang G, "Densely connected convolutional networks", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.4700~4708 (2017)
- Zagoruyko S, "Wide residual networks", arXiv preprint arXiv. 2016;1605.07146, ~ (2016)
- Huang Y, "Gpipe : Efficient training of giant neural networks using pipeline parallelism", Advances in neural information processing systems, 32, pp.103~112 (2019)
- Tan M, "Efficientnet: Rethinking model scaling for convolutional neural networks", International Conference on Machine Learning, pp.6105~6144 (2019), PMLR
- Japkowicz N, "Evaluating learning algorithms: a classification perspective" (2011), Cambridge University Press