LCANet: Learnable Connected Attention Network for Human Identification Using Dental Images.

Area of Science:

• Forensic Science
• Computer Science
• Biomedical Imaging

Background:

• Forensic odontology is crucial for human identification via dental records.
• Accurate matching of dental images remains a challenge in forensic investigations.

Purpose of the Study:

• To develop a novel deep learning method for automated human identification using 2-D panoramic dental X-ray images.
• To improve the accuracy and efficiency of dental image matching in forensic odontology.

Main Methods:

• A novel deep convolution neural network with a top-down architecture was designed.
• The network incorporates an improved channel attention module and a learnable connected module for enhanced feature extraction.
• The method automatically matches 2-D panoramic dental X-ray images for human identification.

Main Results:

• The proposed method achieved new state-of-the-art performance in human identification using dental images.
• Tested on 1,168 images from 503 subjects, the method reached 87.21% rank-1 and 95.34% rank-5 accuracy.
• The channel attention and learnable connected modules improved recognition precision and adaptive layer connections.

Conclusions:

• The novel deep learning approach significantly enhances human identification accuracy in forensic odontology.
• This method offers a promising tool for automated and accurate analysis of dental images for identification purposes.
• The study demonstrates the potential of advanced AI in forensic applications.