Table 1.

Feature extraction. For each work, it is reported whether or not other tasks are performed following feature extraction. The reported results are related to the task following feature extraction. Abbreviations are used for Magnetic Resonance Imaging (MRI), Low Back Pain (LBP), Accuracy (Acc), Mean Absolute Error (MAE), Machine Learning (ML), Support Vector Machine (SVM).

Author/Year	Main Task	Data Type	# Patients	Structures Involved	Results	Model
Adankon, 2012 [19]	Feature Extraction and Classification	3D image of the back surface	165	Vertebrae	Acc = 95%	Local Geometric Descriptors and SVM
Castro-Mateos, 2014 [20]	Feature Extraction and Segmentation	3D MRI	59	Discs	DICE = 88.4%	Statistical shape model space and B-Spline space
Raudner, 2020 [21]	Feature Extraction	MRI	58	Discs	/	GRAPPATINI
Abdollah, 2020 [22]	Feature Extraction	MRI	28	Discs, Vertebrae	/	Random Forest and texture analysis
Yang 2020 [8]	Feature Extraction and Classification	MRI	109	Discs	Acc = 88.3%	Gabor wavelet transformation and KLT feature tracker
Ruiz-España, 2015 [23]	Feature Extraction and Classification	MRI	67	Discs	Acc > 90%	Gradient Vector Flow, several ML models
Ketola, 2020 [24]	Feature Extraction and Classification	MRI	518	LBP	Acc = 83%	Texture feature extraction and Logistic Regression
Garcia-Cano, 2018 [10]	Feature Extraction and Regression	X-rays	150	Vertebrae	Cobb angle MAE = 4.79°	Independent component analysis and Random Forest