Publications by authors named "Andreia Cristina Breda de Souza"
Article Synopsis
- The paper explores the use of Deep Learning (DL) techniques to classify gunshot wounds in a forensic context, focusing on differentiating between entry and exit wounds and identifying the Medical-Legal Shooting Distance (MLSD) based on crime scene images.
- A database of 2,551 images was created, and the ResNet152 architecture outperformed others in classifying gunshot wounds, achieving high accuracy rates of 86.90% for wound classification and 92.48% for MLSD.
- The study highlights the challenges of standardizing forensic images due to varying conditions but emphasizes the potential of DL as a useful tool to support forensic investigations and enhance pathology practices.
Objective: To evaluate macroscopically and radiographically the behavior of four endodontic sealers submitted to high temperatures, with forensic purposes.
Methods: Sixty-four lower anterior teeth were included. All teeth were endodontically treated, and simulation of external cervical resorption (ECR) was performed.