Needle Detection in Ultrasound Images using Ultrasound Beamformed RF Signals

Medical ultrasound imaging is used in many image-guided needle interventions to visualize the patient anatomy and the advancing needle. Despite the improved capability of ultrasound to visualize the patient anatomy, detecting the needle in ultrasound images is often a difficult task. Many methods were introduced to detect the needle in ultrasound images. However, most of these methods employed ultrasound B-mode image analysis to detect the needle. Hence, the accuracy of these methods might be degraded by the limited visibility of the needle in ultrasound B-mode images, particularly at steep insertion angles. This study introduces an effective method to accurately detect the needle in ultrasound images based on the beamformed radio frequency (RF) signals that are recorded during conventional ultrasound imaging procedures. In particular, the beamformed RF signals are processed to identify the needle echoes using two signature-based features. The identified needle echoes are analyzed using the Robust Model Fitting Random Sample Consensus (RANSAC) algorithm to detect the needle trajectory. Moreover, the needle tip is detected using a moving window procedure that quantifies the variations of the beamformed RF signals around the needle trajectory. The performance of the proposed method is evaluated by detecting the trajectories and tips of biopsy needles inserted in bovine liver tissue specimens at different insertion angles. The results show that the proposed method was able to detect the needles with needle angular deviations between 0.2. and 0.6. and needle tip deviations between 0.3 to 0.5 mm. Furthermore, the results indicate that the proposed method was able to outperform a recently introduced needle detection method. These results suggest the feasibility of employing the proposed method to achieve accurate needle detection during ultrasound-guided needle interventions.