Hardware Acceleration Strategies for Mitigating Streak Metal Artifacts in Cone Beam CT: A Survey

Cone Beam Computed Tomography (CBCT) has become increasingly essential in dental, orthopedic, and imageguided surgical applications.However, the presence of metal objects-such as implants or dental fillings-often introduces severe streak artifacts, which degrade image quality and hinder accurate diagnosis.This survey investigates the growing role of hardware acceleration in mitigating these metal-induced artifacts, with a particular focus on strategies that enhance computational efficiency and enable real-time processing.The study begins by reviewing the physical causes of streak artifacts and the limitations of traditional software-based correction techniques.It then categorizes and analyzes a wide range of hardware-accelerated approaches, including GPU-based parallel processing, FPGA implementations, and custom ASIC designs.Special attention is given to algorithm-hardware co-design, where the choice of architecture directly influences the performance of correction algorithms.Comparative evaluation reveals that while GPU solutions offer flexibility and speed, FPGA and ASIC platforms excel in low-power and embedded applications.The review also explores emerging trends such as deep learning-based correction running on specialized hardware.Ultimately, this study highlights that the convergence of hardware acceleration and advanced correction algorithms holds significant promise for delivering artifact-free CBCT imaging, particularly in time-sensitive clinical environments.The paper concludes with current challenges and potential research directions in this fast-evolving domain.