Real-Time Detection and Classification of Porous Bone Structures Using Image Segmentation and Opening Operation Techniques

When treating bone diseases such as osteoarthritis, rheumatoid arthritis, bone fractures, and osteoporosis, the postoperative recovery of patients becomes more efficient by customizing bone implants with different bone densities. In this study, the non-contact surface profile measurement system based on image segmentation and opening operation techniques is proposed to measure the exterior geometric configurations of porous bone implants. We compile the image processing program for distinguishing the bone surface profiles from the original bone images captured by a CMOS camera. First, the original bone images are converted into gray-scale images. Then, the gray-scale images are normalized by k-means clustering for improving the accuracy of binarizations. The opening operation of erosion and dilation is applied to remove image noises and enhance porous features. The Canny edge detection is used to precisely obtain the contours and pixel sizes of the porous bone features. Finally, the clear surface contours of the porous structures without background noises are acquired by the Gaussian blur technique. The geometric dimensions of artificial porous joints and implants can be inspected by this surface profile measurement system after the metal three-dimensional printing process. The experimental results are excessively helpful to real-time implant corrections and accurate treatments.