https://sam.ensam.eu:443 The DSpace digital repository system captures, stores, indexes, preserves, and distributes digital research material. Thu, 14 May 2026 11:09:06 GMT 2026-05-14T11:09:06Z http://hdl.handle.net/10985/15787 Lumbar spine posterior corner detection in X-rays using Haar-based features EBRAHIMI, Shahin; SKALLI, Wafa; ANGELINI, Elsa D.; GAJNY, Laurent 3D reconstruction of the spine using biplanar X-rays remains approximate and requires human-machine interactions to adjust the position of important features such as vertebral corners and endplates. The purpose of this study is to develop a method to extract automatically the accurate position of lumbar vertebrae posterior corners. In the proposed method we select corner point candidates from an initial edge map. A dedicated pipeline is designed to discard unwanted candidates, involving polyline simplification, curvature thresholding and multiscale Haar filtering. Ultimately, we use a priori knowledge derived from an initial 3D spine model to define search areas and select the final corner points. The framework was tested on 21 biplanar X-rays from scoliotic children. Corner positions are compared with manual selections by two experts. The results report a localization accuracy between 0.7 and 1.6 mm, comparable to manual expert variability. Fri, 01 Jan 2016 00:00:00 GMT http://hdl.handle.net/10985/15787 2016-01-01T00:00:00Z EBRAHIMI, Shahin SKALLI, Wafa ANGELINI, Elsa D. GAJNY, Laurent 3D reconstruction of the spine using biplanar X-rays remains approximate and requires human-machine interactions to adjust the position of important features such as vertebral corners and endplates. The purpose of this study is to develop a method to extract automatically the accurate position of lumbar vertebrae posterior corners. In the proposed method we select corner point candidates from an initial edge map. A dedicated pipeline is designed to discard unwanted candidates, involving polyline simplification, curvature thresholding and multiscale Haar filtering. Ultimately, we use a priori knowledge derived from an initial 3D spine model to define search areas and select the final corner points. The framework was tested on 21 biplanar X-rays from scoliotic children. Corner positions are compared with manual selections by two experts. The results report a localization accuracy between 0.7 and 1.6 mm, comparable to manual expert variability. http://hdl.handle.net/10985/15785 Quasi-automatic 3D reconstruction of the full spine from low-dose biplanar X-rays based on statistical inferences and image analysis EBRAHIMI, Shahin; SKALLI, Wafa; ANGELINI, Elsa D.; GAJNY, Laurent; VERGARI, Claudio Purpose: To design a quasi-automated three-dimensional reconstruction method of the spine from biplanar X-rays as the daily used method in clinical routine is based on manual adjustments of a trained operator and the reconstruction time is more than 10 minutes per patient. Methods: The proposed method of 3D reconstruction of the spine (C3-L5) relies first on a new manual input strategy designed to fit clinicians’ skills. Then, a parametric model of the spine is computed using statistical inferences, image analysis techniques and fast manual rigid registration. Results: An agreement study with the clinically used method on a cohort of 57 adolescent scoliotic subjects has shown that both methods have similar performance on vertebral body position and axial rotation (null bias in both cases and standard deviation of signed differences of 1mm and 3.5° around respectively). In average, the solution could be computed in less than 5 minutes of operator time, even for severe scoliosis. Conclusions: The proposed method allows fast and accurate 3D reconstruction of the spine for wide clinical applications and represents a significant step toward full automatization of 3D reconstruction of the spine. Moreover, it is to the best of our knowledge the first method including also the cervical spine. Mon, 01 Jan 2018 00:00:00 GMT http://hdl.handle.net/10985/15785 2018-01-01T00:00:00Z EBRAHIMI, Shahin SKALLI, Wafa ANGELINI, Elsa D. GAJNY, Laurent VERGARI, Claudio Purpose: To design a quasi-automated three-dimensional reconstruction method of the spine from biplanar X-rays as the daily used method in clinical routine is based on manual adjustments of a trained operator and the reconstruction time is more than 10 minutes per patient. Methods: The proposed method of 3D reconstruction of the spine (C3-L5) relies first on a new manual input strategy designed to fit clinicians’ skills. Then, a parametric model of the spine is computed using statistical inferences, image analysis techniques and fast manual rigid registration. Results: An agreement study with the clinically used method on a cohort of 57 adolescent scoliotic subjects has shown that both methods have similar performance on vertebral body position and axial rotation (null bias in both cases and standard deviation of signed differences of 1mm and 3.5° around respectively). In average, the solution could be computed in less than 5 minutes of operator time, even for severe scoliosis. Conclusions: The proposed method allows fast and accurate 3D reconstruction of the spine for wide clinical applications and represents a significant step toward full automatization of 3D reconstruction of the spine. Moreover, it is to the best of our knowledge the first method including also the cervical spine. http://hdl.handle.net/10985/15786 Vertebral corners detection on sagittal X-rays based on shape modelling, random forest classifiers and dedicated visual features EBRAHIMI, Shahin; SKALLI, Wafa; ANGELINI, Elsa D.; GAJNY, Laurent Quantitative measurements of spine shape parameters on planar X-ray images is critical for clinical applications but remains tedious and with no fully-automated solution demonstrated on the whole spine. This study aims to limit manual input, while demonstrating precise vertebrae corners positioning and shape parameter measurements from sagittal radiographs of the cervical and lumbar regions, exploiting novel dedicated visual features and specialized classifiers. A database of manually annotated X-ray images is used to train specialized Random Forest classifiers for each spine regions and corner types. An original combination of local gradient characteristics, Haar-like features, and contextual features based on patch intensity and contrast is used as visual features. The proposed method is evaluated on 49 sagittal X-rays of asymptomatic and pathological subjects, from multiple imaging sites, and with a large age range (6 – 69 years old). Performance is first evaluated for positioning a 2D spine shape model, where precisely detected corners enable to adjust the model via an original multilinear statistical regression. Root-mean square errors (RMSE) of corners localization and vertebra orientations are reported, demonstrating state-of-the-art precision compared to existing methods, but with minimal manual input. The method is then evaluated for the extraction of additional vertebrae shape characteristics, such as centre positioning, endplate centres positioning and endplate length measures, rarely studied in previous literature. The proposed method enables, with minimal initialization, fast and precise individual vertebrae delineations on sagittal radiographs on normal and pathological cases, with a level of precision and robustness required for objective support for diagnosis and therapy decision making. Mon, 01 Jan 2018 00:00:00 GMT http://hdl.handle.net/10985/15786 2018-01-01T00:00:00Z EBRAHIMI, Shahin SKALLI, Wafa ANGELINI, Elsa D. GAJNY, Laurent Quantitative measurements of spine shape parameters on planar X-ray images is critical for clinical applications but remains tedious and with no fully-automated solution demonstrated on the whole spine. This study aims to limit manual input, while demonstrating precise vertebrae corners positioning and shape parameter measurements from sagittal radiographs of the cervical and lumbar regions, exploiting novel dedicated visual features and specialized classifiers. A database of manually annotated X-ray images is used to train specialized Random Forest classifiers for each spine regions and corner types. An original combination of local gradient characteristics, Haar-like features, and contextual features based on patch intensity and contrast is used as visual features. The proposed method is evaluated on 49 sagittal X-rays of asymptomatic and pathological subjects, from multiple imaging sites, and with a large age range (6 – 69 years old). Performance is first evaluated for positioning a 2D spine shape model, where precisely detected corners enable to adjust the model via an original multilinear statistical regression. Root-mean square errors (RMSE) of corners localization and vertebra orientations are reported, demonstrating state-of-the-art precision compared to existing methods, but with minimal manual input. The method is then evaluated for the extraction of additional vertebrae shape characteristics, such as centre positioning, endplate centres positioning and endplate length measures, rarely studied in previous literature. The proposed method enables, with minimal initialization, fast and precise individual vertebrae delineations on sagittal radiographs on normal and pathological cases, with a level of precision and robustness required for objective support for diagnosis and therapy decision making. http://hdl.handle.net/10985/18317 Vertebral rotation estimation from frontal X-rays using a quasi-automated pedicle detection method EBRAHIMI, Shahin; SKALLI, Wafa; ANGELINI, Elsa D.; GAJNY, Laurent; VERGARI, Claudio Purpose Measurement of vertebral axial rotation (VAR) is relevant for the assessment of scoliosis. Stokes method allows estimating VAR in frontal X-rays from the relative position of the pedicles and the vertebral body. This method requires identifying these landmarks for each vertebral level, which is time-consuming. In this work, a quasi-automated method for pedicle detection and VAR estimation was proposed. Method A total of 149 healthy and adolescent idiopathic scoliotic (AIS) subjects were included in this retrospective study. Their frontal X-rays were collected from multiple sites and manually annotated to identify the spinal midline and pedicle positions. Then, an automated pedicle detector was developed based on image analysis, machine learning and fast manual identification of a few landmarks. VARs were calculated using the Stokes method in a validation dataset of 11 healthy (age 6–33 years) and 46 AIS subjects (age 6–16 years, Cobb 10°–46°), both from detected pedicles and those manually annotated to compare them. Sensitivity of pedicle location to the manual inputs was quantified on 20 scoliotic subjects, using 10 perturbed versions of the manual inputs. Results Pedicles centers were localized with a precision of 84% and mean difference of 1.2 ± 1.2 mm, when comparing with manual identification. Comparison of VAR values between automated and manual pedicle localization yielded a signed difference of − 0.2 ± 3.4°. The uncertainty on pedicle location was smaller than 2 mm along each image axis. Conclusion The proposed method allowed calculating VAR values in frontal radiographs with minimal user intervention and robust quasi-automated pedicle localization. Tue, 01 Jan 2019 00:00:00 GMT http://hdl.handle.net/10985/18317 2019-01-01T00:00:00Z EBRAHIMI, Shahin SKALLI, Wafa ANGELINI, Elsa D. GAJNY, Laurent VERGARI, Claudio Purpose Measurement of vertebral axial rotation (VAR) is relevant for the assessment of scoliosis. Stokes method allows estimating VAR in frontal X-rays from the relative position of the pedicles and the vertebral body. This method requires identifying these landmarks for each vertebral level, which is time-consuming. In this work, a quasi-automated method for pedicle detection and VAR estimation was proposed. Method A total of 149 healthy and adolescent idiopathic scoliotic (AIS) subjects were included in this retrospective study. Their frontal X-rays were collected from multiple sites and manually annotated to identify the spinal midline and pedicle positions. Then, an automated pedicle detector was developed based on image analysis, machine learning and fast manual identification of a few landmarks. VARs were calculated using the Stokes method in a validation dataset of 11 healthy (age 6–33 years) and 46 AIS subjects (age 6–16 years, Cobb 10°–46°), both from detected pedicles and those manually annotated to compare them. Sensitivity of pedicle location to the manual inputs was quantified on 20 scoliotic subjects, using 10 perturbed versions of the manual inputs. Results Pedicles centers were localized with a precision of 84% and mean difference of 1.2 ± 1.2 mm, when comparing with manual identification. Comparison of VAR values between automated and manual pedicle localization yielded a signed difference of − 0.2 ± 3.4°. The uncertainty on pedicle location was smaller than 2 mm along each image axis. Conclusion The proposed method allowed calculating VAR values in frontal radiographs with minimal user intervention and robust quasi-automated pedicle localization. http://hdl.handle.net/10985/20157 Automatic Segmentation and Identification of Spinous Processes on Sagittal X-Rays Based on Random Forest Classification and Dedicated Contextual Features EBRAHIMI, Shahin; SKALLI, Wafa; ANGELINI, Elsa D.; GAJNY, Laurent X-ray based quantitative analysis of spine parameters is required in routine diagnosis or treatment planning. Existing tools commonly require manual intervention. Attempts towards automation of the whole procedure have mainly focused on vertebral bodies, whereas other regions such as the posterior arch also bear considerable amount of useful information. In this study, we combine a specific design of contextual visual features with a multi-class Random Forest classifier to perform pixel-wise segmentation and identification of all cervical spine spinous processes, on sagittal radiographs. Segmentations were evaluated on 62 radiographs, comparing to manual tracing. Correct identification was obtained for all subjects, and segmentation returned mean  SD values of: Dice coefficient =88  8%; Hausdorff distance =2.1  1.4 mm and; mean surface distance =0.6  0.4 mm. The derived geometric parameters can be used to reduce the amount of manual intervention needed for spine modeling or to measure clinical indices. Tue, 01 Jan 2019 00:00:00 GMT http://hdl.handle.net/10985/20157 2019-01-01T00:00:00Z EBRAHIMI, Shahin SKALLI, Wafa ANGELINI, Elsa D. GAJNY, Laurent X-ray based quantitative analysis of spine parameters is required in routine diagnosis or treatment planning. Existing tools commonly require manual intervention. Attempts towards automation of the whole procedure have mainly focused on vertebral bodies, whereas other regions such as the posterior arch also bear considerable amount of useful information. In this study, we combine a specific design of contextual visual features with a multi-class Random Forest classifier to perform pixel-wise segmentation and identification of all cervical spine spinous processes, on sagittal radiographs. Segmentations were evaluated on 62 radiographs, comparing to manual tracing. Correct identification was obtained for all subjects, and segmentation returned mean  SD values of: Dice coefficient =88  8%; Hausdorff distance =2.1  1.4 mm and; mean surface distance =0.6  0.4 mm. The derived geometric parameters can be used to reduce the amount of manual intervention needed for spine modeling or to measure clinical indices. http://hdl.handle.net/10985/19796 Quasi-automated reconstruction of the femur from bi-planar X-rays GIRINON, François; EBRAHIMI, Shahin; DAGNEAUX, Louis; SKALLI, Wafa; ROUCH, Philippe; GAJNY, Laurent 3D reconstruction from low-dose Bi-Planar X-Rays (BPXR) is a rising practice in clinical routine. However, this process is time consuming and highly depends on the user. This study aims to partially automate the process for the femur, thus decreasing reconstruction time and increasing robustness. As a training set, 50 femurs are segmented from CT scans together with 120 BPXR reconstructions. From this dataset, an initial solution for the bony contours is defined through Gaussian Process Regression (GPR), using eight digitized landmarks. This initial solution is projected on both x-rays and automatically adjusted using an adapted Minimal Path Algorithm (MPA). To evaluate this method, CT-scans were acquired from 20 cadaveric femurs. For each sample, the CTbased reconstruction is compared to the one automatically generated from the digitally reconstructed radiographs. Euclidean distances between femur reconstructions and the segmented CT data are on average 1.0 mm with a Root Mean Square Error (RMSE) of 0.8 mm. Femoral torsion errors are assessed: the bias is lower than 0.1° with a 95% confidence interval of 4.8°. The proposed method substantially improves 3D reconstructions from BPXR, as it enables a fast and reliable reconstruction, without the need for manual adjustments, which is essential in clinical routine. Wed, 01 Jan 2020 00:00:00 GMT http://hdl.handle.net/10985/19796 2020-01-01T00:00:00Z GIRINON, François EBRAHIMI, Shahin DAGNEAUX, Louis SKALLI, Wafa ROUCH, Philippe GAJNY, Laurent 3D reconstruction from low-dose Bi-Planar X-Rays (BPXR) is a rising practice in clinical routine. However, this process is time consuming and highly depends on the user. This study aims to partially automate the process for the femur, thus decreasing reconstruction time and increasing robustness. As a training set, 50 femurs are segmented from CT scans together with 120 BPXR reconstructions. From this dataset, an initial solution for the bony contours is defined through Gaussian Process Regression (GPR), using eight digitized landmarks. This initial solution is projected on both x-rays and automatically adjusted using an adapted Minimal Path Algorithm (MPA). To evaluate this method, CT-scans were acquired from 20 cadaveric femurs. For each sample, the CTbased reconstruction is compared to the one automatically generated from the digitally reconstructed radiographs. Euclidean distances between femur reconstructions and the segmented CT data are on average 1.0 mm with a Root Mean Square Error (RMSE) of 0.8 mm. Femoral torsion errors are assessed: the bias is lower than 0.1° with a 95% confidence interval of 4.8°. The proposed method substantially improves 3D reconstructions from BPXR, as it enables a fast and reliable reconstruction, without the need for manual adjustments, which is essential in clinical routine.