Surgical instrument tip detection is an important component in Computer-Assisted Laparoscopy, that can be practically applied to tasks such as tracking tool tips, assessing surgeon skills, and more. Current frameworks use training methods that are often computationally expensive. By leveraging available segmentation frameworks, we designed a training-free algorithm for instrument tip detection. Using ground truth segmentation, the algorithm can achieve up to 87% accuracy on the Endovis15 dataset, while using models like Roboflow, YOLOv9c-seg, and YOLOv9e-seg, the average results achieved for 4 videos are 40%, 28.68%, and 31.55%, respectively.