Patterns detection and recognition in visual aided system for prosthesis pose estimation during total hip replacement surgery

Total hip replacement (THR) surgeries are leading to severe complications such as prosthetic impingement and dislocation. To help surgeons during THR surgery we are developing a real-time visual aided system for an accurate placement of hip prostheses within the safe zone. To ensure the feasibility of this visual aided system in the blood interfering situation during surgery, improved pattern detection and recognition method is proposed in this paper. The mini camera mounted on the femoral head is used to take images of customized patterns designed inside the acetabular cup. Firstly, all the blood-covered patterns are detected. Since the blood is red in color for processing, we extract red channel of the image sequence. For noise elimination, edge preservation and uneven illumination, the median filter and adaptive thresholding is applied respectively. Secondly, recognizing each pattern appeared in the frame by generating its specific 9-bit binary, sampling at each pattern from top left corner to bottom right corner. The simulation results show pattern detection and recognition rate as high as 99%, which validates the efficiency of the proposed method.