The Archives of Bone and Joint Surgery

The Archives of Bone and Joint Surgery

Deep Learning-based Detection of Hip Anatomical Landmarks in X-ray Images

Document Type : RESEARCH PAPER

Authors
Mahdie Aghasizade 1
Morad Karimpour 1
Masoud Shariat-Panahi 1
Mohammad Akhlaghi 1
Alireza Almasi Nokiani 2
Ramin Jafarzadeh Khaledi 3
Hadi G. Kashani 1
1 School of Mechanical Engineering, University of Tehran, Tehran, Iran
2 Firoozabadi Clinical Research Development Unit (FACRDU), Iran. University of Medical Sciences (lUMS), Tehran, Iran
3 Rad Radiology and Sonography Clinic, Tehran, Iran
10.22038/abjs.2025.82151.3740
Abstract
Objectives: The accurate quantification of hip deformities in medical images presented a significant challenge for radiologists. Inaccurate quantifications can lead to misclassification of diseases and improper treatment. This study introduced two-stage deep learning (DL) models designed for the detection of landmarks in hip X-ray images, aiming to improve the accuracy of deformity quantification.
Methods: This study employed two two-stage deep learning (DL) models based on VGG16 and ResNet50. The first stage aimed to detect the bounding box for each landmark, while the second stage focused on pinpointing the exact location of the landmark within the magnified box. The model automatically identified 16 hip landmarks and measures 12 specific dimensions on each side. Training, validation, and testing were conducted on a dataset comprising 854 3-joint lower limb X-ray images depicting various anomalies.
Results: The model's measurements were compared with those obtained manually to evaluate performance. The resulting average error ranged from 0.56 to 2.33 mm for different landmarks in the ResNet50-based model. The most accurate measurements were obtained for the femoral head radius and shaft width, with average errors of 0.49 mm and 0.64 mm, respectively. Conversely, the least accurate results were observed for the alpha angle, with an error of 7.37°.
Conclusion: This dataset can be valuable for diverse research endeavors related to the hip region. The 2-stage model demonstrated a brief learning time due to the small image size loaded in both steps, while also offering high precision through the use of a bounding box.
Level of evidence: V
Keywords
Anatomical landmark detection
Deep learning (DL)
Hip abnormalities
Two-stage DL model
Subjects
1.    Wilson JJ, Furukawa M. Evaluation of the patient with hip pain. Am Fam Physician. 2014;89(1):27-34.
2.    Lim SJ, Park YS. Plain radiography of the hip: a review of radiographic techniques and image features. Hip Pelvis. 2015;27(3):125-34. doi: 10.5371/hp.2015.27.3.125.
3.    Kehr P. The hip joint in adults: advances and developments edited by K. Mohan Iyer. European Journal of Orthopaedic Surgery & Traumatology. 2018;28(8):1645-6.
4.    Liu W, Wang Y, Jiang T, Chi Y, Zhang L, Hua XS. Landmarks detection with anatomical constraints for total hip arthroplasty preoperative measurements. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention 2020 (pp. 670-679). Cham: Springer International Publishing.
5.    Torosdagli N, Liberton DK, Verma P, Sincan M, Lee JS, Bagci U. Deep geodesic learning for segmentation and anatomical landmarking. IEEE Trans Med Imaging. 2019;38(4):919-931. doi: 10.1109/TMI.2018.2875814.
6.    Bekkouch IEI, Maksudov B, Kiselev S, Mustafaev T, Vrtovec T, Ibragimov B. Multi-landmark environment analysis with reinforcement learning for pelvic abnormality detection and quantification. Med Image Anal. 2022:78:102417. doi: 10.1016/j.media.2022.102417.
7.    Jo C, Hwang D, Ko S, et al. Deep learning-based landmark recognition and angle measurement of full-leg plain radiographs can be adopted to assess lower extremity alignment. Knee Surg Sports Traumatol Arthrosc.2023;31(4):1388-1397. doi: 10.1007/s00167-022-07124-x.
8.    Juneja M, Garg P, Kaur R, et al. A review on cephalometric landmark detection techniques. Biomedical Signal Processing and Control. 2021;66:102486.
9.    Schwendicke F, Chaurasia A, Arsiwala L, et al. Deep learning for cephalometric landmark detection: systematic review and meta-analysis. Clin Oral Investig. 2021;25(7):4299-4309. doi: 10.1007/s00784-021-03990-w.
10.  Khalid H, Hussain M, Al Ghamdi MA, et al. A comparative systematic literature review on knee bone reports from MRI, X-rays, and CT scans using deep learning and machine learning methodologies. Diagnostics (Basel). 2020;10(8):518. doi: 10.3390/diagnostics10080518.
11.  Hernigou P, Barbier O, Chenaie P. Hip arthroplasty dislocation risk calculator: evaluation of one million primary implants and twenty-five thousand dislocations with deep learning artificial intelligence in a systematic review of reviews. Int Orthop. 2023;47(2):557-571. doi: 10.1007/s00264-022-05644-2.
12.  Liu X, Faes L, Kale AU, et al. A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis. Lancet Digit Health. 2019;1(6):e271-e297. doi: 10.1016/S2589-7500(19)30123-2. 
13.  Kalmet PH, Sanduleanu S, Primakov S, et al. Deep learning in fracture detection: a narrative review. Acta Orthop. 2020;91(2):215-220. doi: 10.1080/17453674.2019.1711323.
14.  Chea P, Mandell JC. Current applications and future directions of deep learning in musculoskeletal radiology. Skeletal Radiol. 2020;49(2):183-197. doi: 10.1007/s00256-019-03284-z.
15.  Fischer MC, Krooß F, Habor J, Radermacher K. A robust method for automatic identification of landmarks on surface models of the pelvis. Sci Rep. 2019;9(1):13322. doi: 10.1038/s41598-019-49573-4.
16.  Thompson P, Medical Annotation Collaborative, Perry DC, Cootes TF, Lindner C. Automation of clinical measurements on radiographs of children’s hips. InInternational Conference on          Medical Image Computing and Computer-Assisted Intervention 2022 (pp. 419-428). Cham: Springer Nature Switzerland.
17.  Liu C, Xie H, Zhang S, Mao Z, Sun J, Zhang Y. Misshapen pelvis landmark detection with local-global feature learning for diagnosing developmental dysplasia of the hip. IEEE Trans Med Imaging. 2020;39(12):3944-3954. doi: 10.1109/TMI.2020.3008382.
18.  McCouat J, Voiculescu I, Glyn-Jones S. Automatically diagnosing hip conditions from x-rays using landmark detection. In2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI) 2021 (pp. 179-182). IEEE.
19.  Ha A, Vorhies J, Campion A, et al. Automatic Extraction of Skeletal Maturity from Whole Body Pediatric Scoliosis X-rays Using Regional Proposal and Compound Scaling Convolutional Neural Networks. In2020 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) 2020 (pp. 996-1000). IEEE.
20.  Wu H, Xie H, Lin F, Zhang S, Sun J, Zhang Y. WaveCSN: cascade segmentation network for hip landmark detection. InProceedings of the 1st ACM International Conference on Multimedia in Asia 2019 (pp. 1-6).
21.  Tack A, Preim B, Zachow S. Fully automated assessment of knee alignment from full-leg X-rays employing a “YOLOv4 And Resnet Landmark regression Algorithm”(YARLA): data from the Osteoarthritis Initiative. Comput Methods Programs Biomed. 2021:205:106080. doi: 10.1016/j.cmpb.2021.106080.
22.  Aghasizade M, Kiyoumarsioskouei A, Hashemi S, et al. A coordinate-regression-based deep learning model for catheter detection during structural heart interventions. Applied Sciences. 2023;13(13):7778.
23.  Archer H, Reine S, Alshaikhsalama A, et al. Artificial intelligence-generated hip radiological measurements are fast and adequate for reliable assessment of hip dysplasia: an external validation study. Bone Jt Open. 2022;3(11):877-884. doi: 10.1302/2633-1462.311.BJO-2022-0125.R1.
24.  Yang W, Ye Q, Ming S, et al. Feasibility of automatic measurements of hip joints based on pelvic radiography and a deep learning algorithm. Eur J Radiol. 2020:132:109303. doi: 10.1016/j.ejrad.2020.109303.
25.  Hussain D, Han S-M, Kim T-S. Automatic hip geometric feature extraction in DXA imaging using regional random forest. J Xray Sci Technol. 2019;27(2):207-236. doi: 10.3233/XST-180434.
26.  Mast NH, Impellizzeri F, Keller S, Leunig M. Reliability and agreement of measures used in radiographic evaluation of the adult hip. Clin Orthop Relat Res. 2011;469(1):188-99. doi: 10.1007/s11999-010-1447-9.
27.  Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9(7):671-5. doi: 10.1038/nmeth.2089.
28.  Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556. 2014 Sep 4.
29.  Grupp RB, Unberath M, Gao C, et al. Automatic annotation of hip anatomy in fluoroscopy for robust and efficient 2D/3D registration. Int J Comput Assist Radiol Surg. 2020;15(5):759-769. doi: 10.1007/s11548-020-02162-7.
30.  Jensen J, Graumann O, Overgaard S, et al. A Deep Learning Algorithm for Radiographic Measurements of the Hip in Adults-A Reliability and Agreement Study. Diagnostics (Basel). 2022;12(11):2597. doi: 10.3390/diagnostics12112597.
31.  Jang SJ, Kunze KN, Vigdorchik JM, Jerabek SA, Mayman DJ, Sculco PK. John Charnley Award: deep learning prediction of hip joint center on standard pelvis radiographs. J Arthroplasty. 2022;37(7S):S400-S407.e1. doi: 10.1016/j.arth.2022.03.033.
32.  Jo C, Hwang D, Ko S, et al. Deep learning-based landmark recognition and angle measurement of full-leg plain radiographs can be adopted to assess lower extremity alignment. Knee Surg Sports Traumatol Arthrosc. 2023;31(4):1388-1397. doi: 10.1007/s00167-022-07124-x.
33.  Nguyen TP, Chae D-S, Park S-J, Kang K-Y, Lee W-S, Yoon J. Intelligent analysis of coronal alignment in lower limbs based on radiographic image with convolutional neural network. Comput Biol Med. 2020:120:103732. doi: 10.1016/j.compbiomed.2020.103732.
34.  Wojciechowski W, Molka A, Tabor Z. Automated measurement of parameters related to the deformities of lower limbs based on x-rays images. Comput Biol Med. 2016:70:1-11. doi: 10.1016/j.compbiomed.2015.12.027.
35.  Cakmak GR, Hamamci IE, Yilmaz MK, Alhajj R, Azboy I, Ozdemir MK. AutoCOR: Autonomous Condylar Offset Ratio Calculator on TKA-Postoperative Lateral Knee X-ray. arXiv preprint arXiv:220403120. 2022;doi:10.2139/ssrn.4856904.
36.  Wu H, Xie H, Liu C, Zha ZJ, Sun J, Zhang Y. Circlenet for hip landmark detection. InProceedings of the AAAI Conference on Artificial Intelligence 2020 (Vol. 34, No. 07, pp. 12370-12377).
37.  Stotter C, Klestil T, Röder C, et al. Deep Learning for Fully Automated Radiographic Measurements of the Pelvis and Hip. Diagnostics (Basel). 2023;13(3):497. doi: 10.3390/diagnostics13030497.
 
 
 
 
 
 
The Archives of Bone and Joint Surgery
Volume 14, Issue 8
August 2026
Pages 527-539

  • Receive Date 16 September 2024
  • Revise Date 27 July 2025
  • Accept Date 04 January 2025