An Accurate Detection Method for Biopsy Needle in Ultrasound Images Based on Fuzzy Enhancement and Hough Transform

Ultrasound guided biopsy is a popular biopsy technology. An important issue for this technology is biopsy needle detection. However, the contrast of ultrasound images is low and the involved background is complicated. Existing needle detection methods cannot realize the fast and accurate needle detection. To address this issue, we have proposed a new needle detection algorithm by combining fuzzy enhancement with Hough transform. In the proposed method, the improved fuzzy enhancement method is firstly utilized to improve the image contrast, and image thinning algorithm is used to extract the linear skeleton. Then, Hough transform is adopted to detect the straight line. Finally, the endpoint detected using Hough transform is corrected based on the intensity correlation of pixels. The proposed method has been evaluated on the ultrasound prostate image sequences including 32 frames where each frame is of size 800x600. Experimental results demonstrate that the proposed method generally has higher detection accuracy than other compared methods in that it can provide 6.7% misdetection rate and its mean length and angle errors are 8.8 pixels and 0.98 degree, respectively. Meanwhile, the implementation time of the proposed method is 25 ms. Indeed, the proposed algorithm can detect the linear surgical instruments accurately and meet the demand of clinical practices.