요추 특징점 추출을 위한 영역 분할 모델의 성능 비교 분석
A Comparative Performance Analysis of Segmentation Models for Lumbar Key-points Extraction
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Most of spinal diseases are diagnosed based on the subjective judgment of a specialist, so numerous stud- ies have been conducted to find objectivity by automating the diagnosis process using deep learning. In this paper, we propose a method that combines segmentation and feature extraction, which are frequently used techniques for diagnosing spinal diseases. Four models, U-Net, U-Net++, DeepLabv3+, and M-Net were trained and compared using 1000 X-ray images, and key-points were derived using Douglas-Peucker algorithms. For evaluation, Dice Sim- ilarity Coefficient(DSC), Intersection over Union(IoU), precision, recall, and area under precision-recall curve eval- uation metrics were used and U-Net++ showed the best performance in all metrics with an average DSC of 0.9724. For the average Euclidean distance between estimated key-points and ground truth, U-Net was the best, followed by U-Net++. However the difference in average distance was about 0.1 pixels, which is not significant. The results suggest that it is possible to extract key-points based on segmentation and that it can be used to accurately diagnose various spinal diseases, including spondylolisthesis, with consistent criteria.
Abstract
Most of spinal diseases are diagnosed based on the subjective judgment of a specialist, so numerous stud- ies have been conducted to find objectivity by automating the diagnosis process using deep learning. In this paper, we propose a method that combines segmentation and feature extraction, which are frequently used techniques for diagnosing spinal diseases. Four models, U-Net, U-Net++, DeepLabv3+, and M-Net were trained and compared using 1000 X-ray images, and key-points were derived using Douglas-Peucker algorithms. For evaluation, Dice Sim- ilarity Coefficient(DSC), Intersection over Union(IoU), precision, recall, and area under precision-recall curve eval- uation metrics were used and U-Net++ showed the best performance in all metrics with an average DSC of 0.9724. For the average Euclidean distance between estimated key-points and ground truth, U-Net was the best, followed by U-Net++. However the difference in average distance was about 0.1 pixels, which is not significant. The results suggest that it is possible to extract key-points based on segmentation and that it can be used to accurately diagnose various spinal diseases, including spondylolisthesis, with consistent criteria.
참고문헌 (18)
- Waxenbaum JA, "Lumbar Vertebrae", StatPerals, pp.29083618~ (2023)
- Nguyen TP, "Deep learning system for Meyerding classification and segmental motion measurement in diagnosis of lumbar spondylolisthesis", Biomedical Signal Processing and Control, 65, pp.102371~ (2021)
- 장준수, "다항 곡선과 합성곱 신경망을 이용한 요추 전위 정량화 방법", 차세대컨버전스정보서비스기술논문지, 10(5), pp.549~557 (2021), 차세대컨버전스정보서비스학회[DOI]
- Afif M, "Deep learning-based technique for lesions segmentation in CT scans images for COVID-19 prediction", Multimedia Tools and Applications, 82, pp.26885~26899 (2023)
- Lu HJ, "Lumbar spine segmentation method based on deep learning", Journal of Applied Clinical Medical Physics, 24(6) (2023)
- Guinebert S, "Automatic semantic segmentation and detection of vertebras and intervertebral discs by neural networks", Computer Methods and Programs in Biomedicine Update, 2, pp.100055~ (2022)
- Ronneberger O, "Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015" (2015), Springer International Publishing
- Mehta R, "M-net: A Convolutional Neural Network for deep brain structure segmentation", 2017 IEEE 14th international symposium on biomedical imaging, pp.437~440 (2017)
- Chen LC, "Computer Vision - ECCV 2018", pp.833~851 (2015), Springer International Publishing
- Zhou Z, "DLMIA ML-CDS 2018", pp.3~11 (2018), Springer International Publishing
- Badrinarayanan V, "SegNet : A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation", IEEE Transactions on Patter Analysis and Machine Intelligence, 39(12), pp.2481~2495 (2017)
- Klinwichit P, "BUU-LSPINE : A Thai Open Lumbar Spine Dataset for Spondylolisthesis Detection", Applied Sciences, 13(15), pp.8646~ (2023)
- Kim DH, "Automated Vertebral Segmentation and Measurement of Vertebral Compression Ratio Based on Deep Learning in X-Ray Images", J Digit Imaging, 34(4), pp.853~861 (2021)
- 백명훈, "딥러닝 기반 척추 자동 분할 및 Cobb angle 측정의 임상적 적용", 차세대컨버전스정보서비스기술논문지, 9(1), pp.1~9 (2020), 차세대컨버전스정보서비스학회[DOI]
- Kim KC, "Automatic detection and segmentation of lumbar vertebrae from X-ray images for compression fracture evaluation", Computer Methods and Programs in Biomedicine, 200, pp.105833~ (2021)
- Douglas DH, "Algorithms for the reduction of the number of points required to represent a digitized line or its caricature", Cartographica : The International Journal for Geographic Information and Geovisualization, 10(2), pp.112~122 (1973)
- Zuiderveld K, "Contrast Limited Adaptive Histogram Equalization. Graphics Gems", pp.474~485 (1994), Academic Press Inc
- Saenpaen J, "A Comparison of Image Enhancement Methods for Lumbar Spine X-ray Image", rison of Image Enhancement Methods for Lumbar Spine X-ray Image. 2018 15th International Conference on Electrical Engineering, 2018, 798~801