Scaphocephaly occurs due to early closure of sagittal suture of skull which results in an abnormal shape of the child's head due to limitation of biparietal width and prominent occipital and frontal bone bossing. Surgical intervention is usually carried out for restoring normal head shape and brain growth. Cranial index is an important tool for determining sagittal suture synostosis severity and effectiveness of surgical intervention. Here, we propose an automatic intelligence-based method for calculating cranial index in affected children. Collected 59 photos (cephalic view) of scaphocephalic patients are firstly preprocessed by Kuwahara filter to enhance the color image quality. Then, Sobel operator is applied to each single channel image for generating edge map. These edge maps (three channels) add to the original image for increasing the contrast of color image. Next, user determines the region of interest (ROI) in the color image, and K-means clustering is applied to separate the head according to color information in the ROI of the input image, generating an output binary image. Finally, cranial index is extracted and compare to hand taken data, measured by pediatric neurosurgeon. Our results show an acceptable accuracy of 94% with fair computational time in comparison with other state-of art methods.
