합성곱 신경망을 이용한 상지 엑스선 영상 분류 모델 유용성 평가
Evaluation of Upper Extremity X-ray Image Classification Software Using Convolution Neural Network
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This study aims to prevent errors that may occur during the radiography examination process, such as misinterpretation of images, by utilizing artificial intelligence, a core technology of the Fourth Industrial Revolution. Through this, we sought to enhance the work efficiency of radiologic technologists, prevent medical accidents. We labeled 19,381 upper ex- tremity musculo-skeletal X-ray images into 7 regions and 19 classes, and divided them into training, validation, and test sets at a ratio of 8:1:1. We used deep learning models such as VGG-16, DenseNet-121, and ResNet-152v2 to evaluate model performance based on accuracy, precision, recall, F1-score, and confusion matrix. The results showed that DenseNet-121 achieved an overall accuracy of 87.77%, an average class accuracy of 98.71%, a precision of 91.78%, a recall of 86.93%, and an F1 score of 86.71%. DenseNet-121 demonstrated the highest performance across all metrics. This study evaluated the performance of various deep learning models using upper extremity radiographic image and demonstrated sufficient performance. Through this, the potential to improve work efficiency and prevent medical accidents was confirmed.
Abstract
This study aims to prevent errors that may occur during the radiography examination process, such as misinterpretation of images, by utilizing artificial intelligence, a core technology of the Fourth Industrial Revolution. Through this, we sought to enhance the work efficiency of radiologic technologists, prevent medical accidents. We labeled 19,381 upper ex- tremity musculo-skeletal X-ray images into 7 regions and 19 classes, and divided them into training, validation, and test sets at a ratio of 8:1:1. We used deep learning models such as VGG-16, DenseNet-121, and ResNet-152v2 to evaluate model performance based on accuracy, precision, recall, F1-score, and confusion matrix. The results showed that DenseNet-121 achieved an overall accuracy of 87.77%, an average class accuracy of 98.71%, a precision of 91.78%, a recall of 86.93%, and an F1 score of 86.71%. DenseNet-121 demonstrated the highest performance across all metrics. This study evaluated the performance of various deep learning models using upper extremity radiographic image and demonstrated sufficient performance. Through this, the potential to improve work efficiency and prevent medical accidents was confirmed.
참고문헌 (25)
- Ker J, "Deep learning applications in medical image analysis", IEEE Access, 6, pp.9375~9389 (2018)
- Hong JY, "Artificial intelligence based medical imaging : An overview", Journal of Radiological Science and Technology, 43(3), pp.195~208 (2020)
- Kiss N, "Mushroom image classification with CNNs : A case-study of different learning strategies", International Symposium on Image and Signal Processing and Analysis(ISPA) (2021)
- Sung J, "Added value of deep learning-based detection system for multiple major findings on chest radiographs : A randomized crossover study", Radiology, 299(2), pp.450~459 (2021)
- Park S, "Artificial intelligence-powered spatial analysis of tumor-infiltrating lymphocytes as complementary biomarker for immune checkpoint inhibition in non-small-cell lung cancer", Journal of Clinical Oncology, 40(17), pp.1916~1928 (2022)
- Larson DB, "Performance of a deep-learning neural network model in assessing skeletal maturity on pediatric hand radiographs", Radiology, 287(1), pp.313~322 (2018)
- McKinney SM, "International evaluation of an AI system for breast cancer screening", Nature, 577(7788), pp.89~94 (2020)
- He K, "Deep residual learning for image recognition", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.770~778 (2016)
- Jeong MJ, "Evaluation of U-net based learning models according to equalization algorithm in thyroid ultrasound imaging", Journal of Radiological Science and Technology, 47(1), pp.29~37 (2024)
- Huang G, "Densely connected convolutional networks", 2017IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.2261~2269 (2017)
- He K, "Deep residual learning for image recognition", 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.770~778 (2016)
- Hong CS, "Confusion plot for the confusion matrix", Journal of the Korean Data And Information Science Society, 32(2), pp.427~437 (2021)[DOI]
- Jeong EH, "Deep learning application of gamma camera quality control in nuclear medicine", Journal of Radiological Science and Technology, 43(6), pp.461~467 (2020)[DOI]
- Yang MJ, "A study on reduced contrast media dose during contrast enhanced abdominal CT : An AI-based contrast enhancement boost", Technique, 26(1), pp.65~73 (2024)
- Kim SH, "Measures to expand the role of radiological technologist in expanding the introduction of artificial intelligence", Journal of Radiation Industry, 13(3), pp.199~206 (2019)[DOI]
- Choi HS, "PACS and teleradiology", Journal of Korean Medical Association, 38(9), pp.1091~1097 (1995)
- Kao CY, "Automated radiology alert system for pneumothorax detection on chest radiographs improves efficiency and diagnostic performance", Diagnostics, 11(7), pp.1182~ (2021)
- Lee HW, "An accuracy evaluation on convolutional neural network assessment of orientation reversal of chest X-ray Image", Journal of Radiological Science and Technology, 43(2), pp.65~70 (2020)
- Kicky G. van L, "How does artificial intelligence in radiology improve efficiency and health outcomes?", Pediatr Radiology (2021)[DOI]
- Reshan MSA, "Detection of pneumonia from chest X-ray images utilizing mobilenet model", Healthcare(Basel), 11(11), pp.1561~ (2023)[DOI]
- Kim KD, "Enhancing deep learning based classifiers with inpainting anatomical side markers(L", R Markers)for multi-center trials, 2022, pp.220, 106705~
- Kicky G. van L, "How does artificial intelligence in radiology improve efficiency and health outcomes?", Pediatr Radiology (2021)[DOI]
- Oh JY, "Convolutional neural network based image correction system for the geometrical error of chest PA X-ray image", ~ (2021)