Lightweight deep learning methods for panoramic dental X-ray image segmentation

Abstract Dental X-ray image segmentation is helpful for assisting clinicians to examine tooth conditions and identify dental diseases. Fast lightweight algorithms without using cloud computing may be required implemented in imaging systems. This paper aims investigate deep learning methods the purpose of deployment on edge devices, such as A novel neural network scheme knowledge distillation proposed this paper. The method a number existing were trained panoramic data set. These evaluated compared by several accuracy metrics. only requires 0.33 million parameters ( $$\sim 7.5$$ ? 7.5 megabytes) model, while it achieved best performance terms IoU (0.804) Dice (0.89) comparing other methods. work shows that small memory storage, comparative performance. could deployed devices potentially assist alleviate their daily workflow improve quality analysis.