Wavelet High-Frequency Enhancement and Frame Dropout Compensation for Accurate Tracking Guidewire/Catheter in Carotid DSA

Digital subtraction angiography (DSA) based interventional surgery is the primary treatment modality for severe carotid artery stenosis (CAS). During the procedure, the 2D position information of guidewire/catheter provided by intraoperative DSA can be registered with the preoperative 3D reconstructed model of the patient's carotid artery, thereby providing realtime 3D visualization of the guidewire/catheter. Therefore, accurate tracking guidewire/catheter in carotid DSA is crucial for ensuring surgical success. However, due to the low contrast of DSA imaging and the potential for occlusion by lesions or other anatomical structures, accurate tracking of the guidewire/catheter remains a challenge. To address these challenges, this paper proposes a method for accurate tracking guidewire/catheter by combining wavelet high-frequency enhancement and frame dropout compensation. Specifically, wavelet enhancement is introduced to improve the detection of sharp high-frequency features, such as the shape contours of guidewire/catheter, and this is combined with interpolation-based compensation for lost frames to achieve continuous, stable, and highly accurate tracking guidewire/catheter in DSA. The proposed method improved the tracking performance of the two superior baseline models by 5.64% and 8.14%, respectively.Clinical relevance - As illustrated in Fig 4, accurate tracking guidewire/catheter during intraoperative DSA, when registered with the patient's preoperative 3D reconstruction model of the carotid artery, provides real-time 3D visualization guidance for the position of the guidewire/catheter to the surgeon.