/LL3rdM-Segmentation-Detection-Extraction

Using deep learning a U-net architecture is used to make segmentation, detection, and extraction of the lower left third molar. The result of the proposed U-net is compared with Attention U-net and U-net++.

Primary LanguageJupyter Notebook

Overview

Using deep learning a U-net architecture is used to make segmentation, detection, and extraction of the lower left third molar. The result of the proposed U-net is compared with Attention U-net and U-net++.

Main Sections

  • Part1: Training a U-net Model for Detecting the lower left third molar and compare it with additional developed models U-net++ and attention U-net .
  • Part2: Models Prediction.
  • Part3: Models Evaluation.

Lower Left Third Molar (LL3rdM) Detection

Three models are evaluated:
  • U-net
  • Attention U-net
  • U-net++

Proposed U-net Model Architecture

Segmentation Results