Deep learning-based human gunshot wounds classification

被引:4
作者
Lira, Renato Queiroz Nogueira [1 ]
de Sousa, Luana Geovana Motta [1 ]
Pinho, Maisa Luana Memoria [1 ]
de Lima, Renan Cesar Pinto da Silva Andrade [1 ]
Freitas, Pedro Garcia [2 ]
Dias, Bruno Scholles Soares [3 ]
de Souza, Andreia Cristina Breda [4 ]
Leite, Andre Ferreira [1 ]
机构
[1] Univ Brasilia, Dept Dent, Campus Darcy Ribeiro, BR-70910900 Brasilia, DF, Brazil
[2] Univ Brasilia, Dept Comp Sci, Campus Darcy Ribeiro, BR-70910900 Brasilia, DF, Brazil
[3] Univ Brasilia, Dept Elect Engn, Campus Darcy Ribeiro, BR-70910900 Brasilia, DF, Brazil
[4] Univ Fed Rio de Janeiro, Dept Forens Dent & Publ Hlth, Cidade Univ, BR-21941617 Rio De Janeiro, RJ, Brazil
关键词
Forensic medicine; Gunshot wounds; Artificial intelligence; Deep learning; SKIN;
D O I
10.1007/s00414-024-03355-4
中图分类号
DF [法律]; D9 [法律]; R [医药、卫生];
学科分类号
0301 ; 10 ;
摘要
In this paper, we present a forensic perspective on classifying gunshot wound patterns using Deep Learning (DL). Although DL has revolutionized various medical specialties, such as automating tasks like medical image classification, its applications in forensic contexts have been limited despite the inherently visual nature of the field. This study investigates the application of DL techniques (59 architectures) to classify gunshot wounds in a forensic context, focusing on distinguishing between entry and exit wounds and determining the Medical-Legal Shooting Distance (MLSD), which classifies wounds as contact, close range, or distant, based on digital images from real crime scene cases. A comprehensive database was constructed with 2,551 images, including 1,883 entries and 668 exit wounds. The ResNet152 architecture demonstrated superior performance in both entry and exit wound classification and MLSD categorization. For the first task, achieved accuracy of 86.90% and an AUC of 82.09%. For MLSD, the ResNet152 showed an accuracy of 92.48% and AUC up to 94.36%, though sample imbalance affected the metrics. Our findings underscore the challenges of standardizing wound images due to varying capture conditions but reflect the practical realities of forensic work. This research highlights the significant potential of DL in enhancing forensic pathology practices, advocating for Artificial Intelligence (AI) as a supportive tool to complement human expertise in forensic investigations.
引用
收藏
页码:651 / 666
页数:16
相关论文
共 50 条
[41]   Deep Learning-Based Phenological Event Modeling for Classification of Crops [J].
Arun, Pattathal V. ;
Karnieli, Arnon .
REMOTE SENSING, 2021, 13 (13)
[42]   A Hybrid Deep Learning-Based Approach for Brain Tumor Classification [J].
Raza, Asaf ;
Ayub, Huma ;
Khan, Javed Ali ;
Ahmad, Ijaz ;
Salama, Ahmed S. ;
Daradkeh, Yousef Ibrahim ;
Javeed, Danish ;
Rehman, Ateeq Ur ;
Hamam, Habib .
ELECTRONICS, 2022, 11 (07)
[43]   Deep Learning-Based ECG Arrhythmia Classification: A Systematic Review [J].
Xiao, Qiao ;
Lee, Khuan ;
Mokhtar, Siti Aisah ;
Ismail, Iskasymar ;
Pauzi, Ahmad Luqman bin Md ;
Zhang, Qiuxia ;
Lim, Poh Ying .
APPLIED SCIENCES-BASEL, 2023, 13 (08)
[44]   Deep Learning-Based Gender Classification by Training With Fake Data [J].
Oulad-Kaddour, Mohamed ;
Haddadou, Hamid ;
Vilda, Cristina Conde ;
Palacios-Alonso, Daniel ;
Benatchba, Karima ;
Cabello, Enrique .
IEEE ACCESS, 2023, 11 :120766-120779
[45]   Data Augmentation for Deep Learning-Based Radio Modulation Classification [J].
Huang, Liang ;
Pan, Weijian ;
Zhang, You ;
Qian, Liping ;
Gao, Nan ;
Wu, Yuan .
IEEE ACCESS, 2020, 8 :1498-1506
[46]   Analyzing the Performance of Deep Learning-based Techniques for Human Pose Estimation [J].
Boscolo, Federico ;
Lamberti, Fabrizio ;
Morra, Lia .
2024 IEEE INTERNATIONAL WORKSHOP ON SPORT, TECHNOLOGY AND RESEARCH, STAR 2024, 2024, :193-198
[47]   Explainable Deep Learning-Based Approach for Multilabel Classification of Electrocardiogram [J].
Ganeshkumar, M. ;
Ravi, Vinayakumar ;
Sowmya, V. ;
Gopalakrishnan, E. A. ;
Soman, K. P. .
IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, 2023, 70 (08) :2787-2799
[48]   Deep Learning-Based Attack Detection and Classification in Android Devices [J].
Gomez, Alfonso ;
Munoz, Antonio .
ELECTRONICS, 2023, 12 (15)
[49]   Deep Learning-Based Classification of Raw Hydroacoustic Signal: A Review [J].
Lin, Xu ;
Dong, Ruichun ;
Lv, Zhichao .
JOURNAL OF MARINE SCIENCE AND ENGINEERING, 2023, 11 (01)
[50]   Deep learning-based classification model for botnet attack detection [J].
Abdulghani Ali Ahmed ;
Waheb A. Jabbar ;
Ali Safaa Sadiq ;
Hiran Patel .
Journal of Ambient Intelligence and Humanized Computing, 2022, 13 :3457-3466