4.2 Article

Automated wound segmentation and classification of seven common injuries in forensic medicine

Journal

Publisher

HUMANA PRESS INC
DOI: 10.1007/s12024-023-00668-5

Keywords

Deep learning; Forensic sciences; Image segmentation; Wound classification

Ask authors/readers for more resources

In forensic medical investigations, automatic segmentation and classification of physical injuries documented with photographs could improve assessment and accelerate reporting. This pilot study trained and compared preexisting deep learning architectures for image segmentation and wound classification. The best models achieved a mean pixel accuracy of 69.4% and mean intersection over union (IoU) of 48.6% on the test set, although they struggled with distinguishing background from wounded areas. Despite difficulties with undefined wound boundaries, the best trained models reliably distinguished among seven common wounds encountered in forensic medical investigations.
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.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.2
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available