Project Overview
Our CBCT super-resolution research addresses a fundamental limitation in dental imaging where cone-beam computed tomography (CBCT) provides insufficient detail for precise endodontic diagnosis and treatment planning. The project develops advanced deep learning models that enhance CBCT image quality using micro-CT (μCT) scans as high-resolution ground truth references, enabling the system to learn complex mappings between clinical CBCT data and superior μCT anatomical detail. The core technical innovation combines convolutional neural networks with specialized denoising algorithms to correct tomography artifacts, enhance resolution, and restore fine anatomical features such as root canal systems, periapical lesions, and bone margins. Key challenges include handling the significant resolution and field-of-view differences between CBCT and μCT modalities, developing robust artifact correction for beam hardening and scatter effects, ensuring anatomical accuracy preservation during enhancement, and achieving real-time processing compatible with clinical workflows.
This breakthrough technology transforms endodontic practice by providing clinicians with dramatically improved image quality for complex root canal diagnoses, surgical planning, and treatment outcome assessment. The enhanced images enable more confident identification of anatomical variations, pathological conditions, and treatment complications that are often missed in standard CBCT scans, ultimately leading to better patient outcomes and reduced treatment failures. Clinical applications extend beyond endodontics to implant planning, oral surgery, and orthodontic analysis where superior image quality directly impacts treatment success. Our research team actively collaborates with dental schools for clinical validation, imaging companies for commercial integration, and endodontic practices for real-world testing, seeking partners ready to implement cutting-edge AI technology that will establish new standards in dental imaging and diagnosis.