دورية أكاديمية
Automated wound segmentation and classification of seven common injuries in forensic medicine.
| العنوان: | Automated wound segmentation and classification of seven common injuries in forensic medicine. |
|---|---|
| المؤلفون: | Zimmermann N; Zurich Institute of Forensic Medicine, University of Zurich, Winterthurerstrasse 190/52, CH-8057, Zurich, Switzerland., Sieberth T; Zurich Institute of Forensic Medicine, University of Zurich, Winterthurerstrasse 190/52, CH-8057, Zurich, Switzerland.; 3D Centre Zurich, University of Zurich, Winterthurerstrasse 190/52, CH-8057, Zurich, Switzerland., Dobay A; Zurich Institute of Forensic Medicine, University of Zurich, Winterthurerstrasse 190/52, CH-8057, Zurich, Switzerland. akos.dobay@uzh.ch.; Forensic Machine Learning Technology Center, University of Zurich, Winterthurerstrasse 190/52, CH-8057, Zurich, Switzerland. akos.dobay@uzh.ch. |
| المصدر: | Forensic science, medicine, and pathology [Forensic Sci Med Pathol] 2024 Jun; Vol. 20 (2), pp. 443-451. Date of Electronic Publication: 2023 Jun 28. |
| نوع المنشور: | Journal Article |
| اللغة: | English |
| بيانات الدورية: | Publisher: Humana Press Country of Publication: United States NLM ID: 101236111 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1556-2891 (Electronic) Linking ISSN: 1547769X NLM ISO Abbreviation: Forensic Sci Med Pathol Subsets: MEDLINE |
| أسماء مطبوعة: | Original Publication: Totowa, N.J. : Humana Press, c2005- |
| مواضيع طبية MeSH: | Photography* , Forensic Medicine*/methods , Wounds and Injuries*/classification , Wounds and Injuries*/diagnostic imaging , Wounds and Injuries*/pathology , Hematoma*/diagnostic imaging , Hematoma*/pathology, Humans ; Pilot Projects ; Deep Learning ; Image Processing, Computer-Assisted |
| مستخلص: | In forensic medical investigations, physical injuries are documented with photographs accompanied by written reports. Automatic segmentation and classification of wounds on these photographs could provide forensic pathologists with a tool to improve the assessment of injuries and accelerate the reporting process. In this pilot study, we trained and compared several preexisting deep learning architectures for image segmentation and wound classification on forensically relevant photographs in our database. The best scores were a mean pixel accuracy of 69.4% and a mean intersection over union (IoU) of 48.6% when evaluating the trained models on our test set. The models had difficulty distinguishing the background from wounded areas. As an example, image pixels showing subcutaneous hematomas or skin abrasions were assigned to the background class in 31% of cases. Stab wounds, on the other hand, were reliably classified with a pixel accuracy of 93%. These results can be partially attributed to undefined wound boundaries for some types of injuries, such as subcutaneous hematoma. However, despite the large class imbalance, we demonstrate that the best trained models could reliably distinguish among seven of the most common wounds encountered in forensic medical investigations. (© 2023. The Author(s).) |
| References: | IEEE Trans Med Imaging. 2010 Feb;29(2):410-27. (PMID: 19825516) Sci Rep. 2020 Dec 14;10(1):21897. (PMID: 33318503) Sensors (Basel). 2020 May 21;20(10):. (PMID: 32455753) J Burn Care Res. 2021 Aug 4;42(4):755-762. (PMID: 33336696) Adv Wound Care (New Rochelle). 2022 Dec;11(12):687-709. (PMID: 34544270) Int J Legal Med. 2021 Sep;135(5):2101-2106. (PMID: 33821334) Burns Trauma. 2019 Feb 28;7:6. (PMID: 30859107) IEEE Access. 2020;8:181590-181604. (PMID: 33251080) Forensic Sci Int. 2018 Jul;288:46-52. (PMID: 29715622) Annu Int Conf IEEE Eng Med Biol Soc. 2015;2015:2415-8. (PMID: 26736781) Comput Intell Neurosci. 2018 May 31;2018:4149103. (PMID: 29955227) Comput Math Methods Med. 2021 Apr 7;2021:5514224. (PMID: 33880130) IEEE Trans Pattern Anal Mach Intell. 2022 Jul;44(7):3523-3542. (PMID: 33596172) PLoS One. 2022 Feb 17;17(2):e0264139. (PMID: 35176101) |
| فهرسة مساهمة: | Keywords: Deep learning; Forensic sciences; Image segmentation; Wound classification |
| تواريخ الأحداث: | Date Created: 20230628 Date Completed: 20240802 Latest Revision: 20240805 |
| رمز التحديث: | 20240805 |
| مُعرف محوري في PubMed: | PMC11297066 |
| DOI: | 10.1007/s12024-023-00668-5 |
| PMID: | 37378809 |
| قاعدة البيانات: | MEDLINE |
Find this article in full text from Springer Verlag. 
Full Text Finder | تدمد: | 1556-2891 |
|---|---|
| DOI: | 10.1007/s12024-023-00668-5 |