Evaluation of scoliosis x-ray images using MATLAB with a database integrated GUI

• Main Authors: Aggabao, Anthony Jude Hermeroll Q., Ong, John Gilbert L., Cadiente, Francis Maynard P., Lim, John Christopher T., Silva, Nigel A.
• Format: text
• Language: English
• Published: Animo Repository 2017
• Subjects: Lumbar curve; Diagnosis, Radioscopic; Scoliosis; Spine > Abnormalities
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/8209
• Institution: De La Salle University

This thesis aims [to] create a program that is able to create evaluations of scoliosis that is close to that of the doctors' evaluations. An alternative system was created that is simple to use and that it makes the diagnosis of potential scoliosis patients easier compare to the standard convention. The images are in DICOM format. These images are in the standard anterior posterior x-ray image for the Cobb angle samples. The lumbar samples on the other hand are samples that display the pelvis along with the lumber area of the spine. These images are processed using the program created in MATLAB to extract the evaluation of the scoliosis patients. For the Cobb samples, the Cobb angle of the patient is extracted using the program and using the points chosen by the program the location of the curve is also determined. For the lumbar samples, the program classifies the samples using the lumbar modifier definition which is based on the lenke classification system which is a standard used in the medical field. The evaluation are exported in an excel file. For Scoliosis Detection, a 98.55 percent passing rate was achieved (68/69) samples. For Cobb Angles with +-5 degrees tolerance from the doctor's measurements, an 88.14 percent passing rate was achieved (52/59) for patients having scoliosis. The algorithm for classifying the lumbar modifier as well as the other parameters were checked and certi[fi]ed by 2 doctors.