Lenkes Scoliosis Classification Using Image Processing Techniques

Abstract: One of major spinal deformities is scoliosis and is diagnosed using posterior-anterior radiograph. The diagnosis and subsequent treatment of the problem depends on the type of scoliosis. Classification by radiograph measure is used to determine the extent of arthrodesis. Although the Kings classification technique has been used more widely because of its simplicity; it is not as effective a method of classification as the Lenkes classification because of low intra observer and inter-observer reliability. Various aspects of the process to automate the efficient classification of Scoliosis radiographs on the basis of Lenkes system using image processing techniques is presented in this paper. Effective image segmentation techniques are used to extract the spinal boundary and inter-vertebral boundaries with help of different image processing software. These shall, in turn, be used in a formulated algorithm to determine and classify the type of scoliosis seen in a radiograph, based on Lenke’s classification system. Objective measurement by the proposed method can eliminate some sources of inter and intra observer error in classification of scoliosis.

Keywords: Scoliosis, Lenke’s Classification, Scoliosis classificaion