A Novel Approach towards Early Detection of Obliteration in Lumbar Lordosis.

Abstract

The millennial age group (18 to 30 years) spend at least 6 hours sitting, either in college or at their workspace. High screen time as a routine, is the major cause for numerous spinal problems. Despite the wide research carried out on postural abnormalities, there exists numerous unrequited queries with regards to lumbar lordosis estimations, due to indeterminate parameters such as age, gender, lifestyle and diet. This work emphasizes the proficient method by observing the posture of a person for early detection of obliteration in Lumbar Lordosis. This further contributes to efficient diagnosis and treatment of spine ailments. With a novel approach to hardware using the myRIO hardware coupled with LabVIEW for interactive interface, the calibration is enhanced using machine learning (ML) - kNN Classifier. The use of machine learning accounts for the variations in the ideal angles of segmented sagittal measures with respect to different subjects. The device is developed to be a non-invasive, user friendly instrument to analyse the casual seated posture trends of the subject. The male subjects are expected to show the tilt angles in the range of -16.3 to -17.2 degrees and similar threshold for females are -15.8 to -16.8 degrees. Out of 120 subjects taken into consideration, the device could accurately classify subjects with obliterated or normal lumbar lordosis). An accuracy and f1- score of 94% and 90% respectively was achieved by the ML model.