Title: Deep learning-based gender classification with dental X-ray images

Authors: B. Vijayakumari; S. Vidhya; J. Saranya

Addresses: Department of ECE, Mepco Schlenk Engineering College, Sivakasi, Tamilnadu, India ' Department of ECE, Mepco Schlenk Engineering College, Sivakasi, Tamilnadu, India ' Department of ECE, Mepco Schlenk Engineering College, Sivakasi, Tamilnadu, India

Abstract: In a forensic department, teeth play a crucial role to recognise a dead or missing person. In forensic analysis, gender difference is a considerable one. Yet, gender identification with dental images using deep learning methods are still in research. An algorithm is proposed here to find human gender using panoramic dental X-ray images (DXI). This work is organised into three sections such as image pre-processing, gradient-based recursive threshold (GBRT) segmentation and classification. Initially, prime magic square filter is used to remove the unwanted noises. Secondly, segmentation GBRT is used. Finally with Resnet50 network, the gender is classified. The dataset of 285 dental images were taken and they are augmented to 4,000 dental images and then they are separated as 3,000 images for training and 1,000 images for testing to carry out experimental evaluation. It provides classification accuracy of 94%. It shows that the proposed work gives convincing results.

Keywords: gender classification; dental radiographs; morphological operations; GBRT segmentation; deep CNN ResNet50 classified results.

DOI: 10.1504/IJBET.2023.131694

International Journal of Biomedical Engineering and Technology, 2023 Vol.42 No.1, pp.109 - 121

Received: 27 Apr 2022
Accepted: 05 Jul 2022
Published online: 27 Jun 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article