Estimation of Spinal Curvature Using Machine Learning

1S.G. Shivabhinav, K.S. Puvivarun and J. Briskilal


One of the common types of scoliosis is Adolescent Idiopathic Scoliosis which affects children between ages 10 to 18.Generally, AIS curves progresses rapidly during the teenage years of patients. Growth progression of many curves grow significantly during skeletal maturity, but curves with more than 60o, progresses even during adult-hood. Since this problem it starts at an early age is difficult to diagnose it since the angle of curvature is small and is generally recognised at older ages say around 45 years. Symptoms of this is generally not observed at the early age and generally is visible during late teens. It causes lower back pain, height asymmetry, lean torso, and may cause problems to nerves. To solve this problem generally we measure Cobb angle manually which is more time-consuming and unreliable. It is very challenging to achieve a highly accurate estimation of Cobb angles as it is difficult to utilize the information of x-rays efficiently. This has sparked interest in developing methods for accurate automated spinal curvature estimation and error correction in spinal anterior-posterior x-ray images. This is done using convolutional neural networks (CNN) and other artificial neural networks (ANN).In order to estimate accurate spinal curvature and Cobb angle we use machine learning.


AIS-Adolescent Idiopathic Scoliosis, Cobb Angle, CNN-Convolutional Neural Networks, ANNArtificial Neural Networks.

Paper Details
IssueIssue 5