A Method of Hand Segmentation by Integrated Image Processing Techniques for Touchless-Controlled Image Viewer in an Operating Room

Sterilization condition in the surgical units is important. Contamination problems regards to infection rate after surgery. The computer workstation is always used for x-ray image viewer. To avoid the contamination problem, we proposed a method of hand segmentation using integrated image processing technique. This is important step to develop a touchless system to control image viewer in an operating room. The method consists of three steps: 1) hand localization, 2) hand detection, and 3) hand segmentation. Hand localization is to use multiple image processing technique for preliminary detecting the hand location. The edge of object inside the hand location is used to define the bounding box in hand detection step. Finally, the hand segmentation is to fill the blank pixel in the hand area. To evaluate the proposed method, data of nine subjects was used. Segmentation precision is measured by the Intersection over Union (IoU) threshold and the detection rate. The detection rate is 100%. The experiment shows satisfactory result.