Skip to main content

Advertisement

SpringerLink
Log in

Human identification performed with skull’s sphenoid sinus based on deep learning

  • Original Article
  • Published:
International Journal of Legal Medicine Aims and scope Submit manuscript

Abstract

Human identification plays a significant role in the investigations of disasters and criminal cases. Human identification could be achieved quickly and efficiently via 3D sphenoid sinus models by customized convolutional neural networks. In this retrospective study, a deep learning neural network was proposed to achieve human identification of 1475 noncontrast thin-slice CT scans. A total of 732 patients were retrieved and studied (82% for model training and 18% for testing). By establishing an individual recognition framework, the anonymous sphenoid sinus model was matched and cross-tested, and the performance of the framework also was evaluated on the test set using the recognition rate, ROC curve and identification speed. Finally, manual matching was performed based on the framework results in the test set. Out of a total of 732 subjects (mean age 46.45 years ± 14.92 (SD); 349 women), 600 subjects were trained, and 132 subjects were tested. The present automatic human identification has achieved Rank 1 and Rank 5 accuracy values of 93.94% and 99.24%, respectively, in the test set. In addition, all the identifications were completed within 55 s, which manifested the inference speed of the test set. We used the comparison results of the MVSS-Net to exclude sphenoid sinus models with low similarity and carried out traditional visual comparisons of the CT anatomical aspects of the sphenoid sinus of 132 individuals with an accuracy of 100%. The customized deep learning framework achieves reliable and fast human identification based on a 3D sphenoid sinus and can assist forensic radiologists in human identification accuracy.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3

source code is open to access. Conv: convolution layer. BN: batchnorm layer. ReLU: activation layer (relu-activation). Max pool: max-pooling layer. Global ave pool: global average pooling layer

Fig. 4
Fig. 5

Similar content being viewed by others

References

  1. Disaster Victim Identification (DVI). International Criminal Police Organization.https://www.interpol.int/en/How-we-work/Forensics/Disaster-Victim-Identification-DVI. Accessed 26 Sept 2021.

  2. Culbert WL, LAW, F. M. (1927) Identification by comparison of roentgenograms: of nasal accessory sinuses and mastoid processes. J Am Med Assoc 88(21):1634–1636

    Article  Google Scholar 

  3. Yoshino M et al (1987) Classification system of frontal sinus patterns by radiography. Its application to identification of unknown skeletal remains. Forensic Sci Int 34(4):289–99

    Article  CAS  Google Scholar 

  4. Tatlisumak E et al (2007) Identification of unknown bodies by using CT images of frontal sinus. Forensic Sci Int 166(1):42–48

    Article  Google Scholar 

  5. Gibelli, Daniele, et al. "Relationship between sphenoid sinus volume and accessory septations: a 3D assessment of risky anatomical variants for endoscopic surgery." The Anatomical Record 303.5 (2020): 1300–1304.

  6. Gibelli, Daniele, et al. "Relationship between sphenoid sinus volume and protrusion of internal carotid artery and optic nerve: a 3D segmentation study on maxillofacial CT-scans." Surgical and Radiologic Anatomy 41.5 (2019): 507–512.

  7. Beaini TL et al (2015) Human identification through frontal sinus 3D superimposition: pilot study with cone beam computer tomography. J Forensic Leg Med 36:63–69

    Article  Google Scholar 

  8. Auffret M, Garetier M, Diallo I, Aho S, Ben Salem D (2016) Contribution of the computed tomography of the anatomical aspects of the sphenoid sinuses to forensic identification. J Neuroradiol 43:404–414

    Article  Google Scholar 

  9. Deloire L et al (2019) Post-mortem X-ray computed tomography (PMCT) identification using ante-mortem CT-scan of the sphenoid sinus. J Neuroradiol 46(4):248–255

    Article  Google Scholar 

  10. Dedouit F et al (2007) Virtual anthropology and forensic identification: report of one case. Forensic Sci Int 173(2):182–187

    Article  Google Scholar 

  11. Ruder TD et al (2011) Radiologic identification of disaster victims: a simple and reliable method using CT of the paranasal sinuses. Eur J Radiol 81(2):e132–e138

    Article  Google Scholar 

  12. Cappella A, Gibelli D, Cellina M et al (2019) Three-dimensional analysis of sphenoid sinus uniqueness for assessing personal identification: a novel method based on 3D–3D superimposition. Int J Legal Med 133:1895–1901. https://doi.org/10.1007/s00414-019-02139-5

    Article  PubMed  Google Scholar 

  13. Cicero M et al (2016) Training and validating a deep convolutional neural network for computer-aided detection and classification of abnormalities on frontal chest radiographs. Invest Radiol 52(5):281–287

    Article  Google Scholar 

  14. Gonzalez G et al (2018) Disease staging and prognosis in smokers using deep learning in chest computed tomography. Am J Respir Crit Care Med 197(2):193–203

    Article  PubMed Central  Google Scholar 

  15. Deng W et al (2019) Brain tumor segmentation based on improved convolutional neural network in combination with non-quantifiable local texture feature. J Med Syst 43(6):152

    Article  Google Scholar 

  16. Spampinato C et al (2017) Deep learning for automated skeletal bone age assessment in X-ray images. Med Image Anal 36:41–51

    Article  CAS  Google Scholar 

  17. Li Y et al (2019) Forensic age estimation for pelvic X-ray images using deep learning. Eur Radiol 29(5):2322–2329

    Article  Google Scholar 

  18. Miki Y et al (2017) Classification of teeth in cone-beam CT using deep convolutional neural network. Comput Biol Med 80:24–29

    Article  Google Scholar 

  19. Zhou Y et al (2019) Digital whole-slide image analysis for automated diatom test in forensic cases of drowning using a convolutional neural network algorithm. Forensic Sci Int 302:109922

    Article  Google Scholar 

  20. Souadih K, Belaid A, Salem DB, Conze PH (2020) Automatic forensic identification using 3D sphenoid sinus segmentation and deep characterization. Med Biol Eng Compu 58(2):291–306

    Article  Google Scholar 

  21. Fastuca, R., Lorusso, P., Lagravère, M. O., Michelotti, A., Portelli, M., Zecca, P. A., ... & Caprioglio, A. (2017). Digital evaluation of nasal changes induced by rapid maxillary expansion with different anchorage and appliance design. BMC Oral Health, 17(1), 1-7.

  22. Hang Su, Subhransu Maji, Evangelos Kalogerakis, Erik Learned-Miller, "Multi-view convolutional neural networks for 3D shape recognition", Proceedings of ICCV 2015

  23. Wang J et al (2010) Extensions of the sphenoid sinus: a new classification. Neurosurgery 66(4):797

    Article  Google Scholar 

  24. Li S, Wang Z, Xian J (2010) Study of variations in adult sphenoid sinus by multislice spiral computed tomography. Zhonghua Yi Xue Za Zhi 90(31):2172

    PubMed  Google Scholar 

  25. Tomovic S et al (2013) High-resolution computed tomography analysis of variations of the sphenoid sinus. J Neurol Surg B Skull Base 74(2):82–90

    Article  PubMed Central  Google Scholar 

  26. Štoković N et al (2016) Sphenoid sinus types, dimensions and relationship with surrounding structures. Annals of Anatomy - Anatomischer Anzeiger 203:69–76

    Article  Google Scholar 

  27. Unal B et al (2006) Risky anatomic variations of sphenoid sinus for surgery. Surg Radiol Anat 28(2):195–201

    Article  Google Scholar 

  28. Ruder TD et al (2016) Comparative radiologic identification with CT images of paranasal sinuses – development of a standardized approach. Journal of Forensic Radiology and Imaging 7:1–9

    Article  Google Scholar 

  29. Brun CN et al (2017) Comparative radiologic identification with standardized single CT images of the paranasal sinuses—evaluation of inter-rater reliability. Forensic Sci Int 280:81–86

    Article  Google Scholar 

  30. Bewes J et al (2019) Artificial intelligence for sex determination of skeletal remains: application of a deep learning artificial neural network to human skulls. J Forensic Leg Med 62:40–43

    Article  Google Scholar 

  31. Nassar DEM, Ammar HH (2007) A neural network system for matching dental radiographs. Pattern Recogn 40(1):65–79

    Article  Google Scholar 

Download references

Author information

Author notes

  1. Hu Chen and Zhenhua Deng equally contributed to the work

Authors and Affiliations

  1. College of Computer Science, Sichuan University, Chengdu, 610065, People’s Republic of China

    Hanjie Wen, Hu Chen, Yi Zhang & Weiqiang Lv

  2. West China School of Basic Medical Sciences & Forensic Medicine, Sichuan University, Chengdu, 610041, People’s Republic of China

    Wei Wu, Fei Fan & Zhenhua Deng

  3. Department of Scientific Research and Education, The Sixth People’s 3. Hospital of Chengdu, Chengdu, 610065, People’s Republic of China

    Peixi Liao

Authors
  1. Hanjie Wen
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Wei Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Fei Fan
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Peixi Liao
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Hu Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Yi Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Zhenhua Deng
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Weiqiang Lv
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Hu Chen or Zhenhua Deng.

Ethics declarations

Ethics approval

This was a retrospective study; the study sample was obtained from previously available medical materials, in order to avoid unnecessary or additional radiation exposure to the patients. The present study was performed with the approval of the ethics committee of the West China Second University Hospital of Sichuan University, and the code of the ethical approval is KS2021507. Informed consent was waived because this was a retrospective study.

Conflict of interest

The authors declare no competing interests.

Additional information

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wen, H., Wu, W., Fan, F. et al. Human identification performed with skull’s sphenoid sinus based on deep learning. Int J Legal Med 136, 1067–1074 (2022). https://doi.org/10.1007/s00414-021-02761-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00414-021-02761-2

Keywords

Search

Navigation