Identification of Postural Deviations and Trunk Asymmetries in Physical Therapy Students Using an AI-Based Mobile Application

The aim of the study was to identify postural deviations and trunk asymmetries by comparing the anterior and posterior body views in physical therapy students using a mobile application based on artificial intelligence. The study included 49 students, specializing in physical therapy, aged between 21-24 years. The evaluation was carried out with the mobile application AI Posture Evaluation and Correction System (APECS), which uses markers placed on the body photograph and photogrammetric algorithms for precise scanning of body symmetry. The analysis included 6 transversal parameters and 3 trunk symmetry indices. The analysis of the results of body posture differences highlights that in the anterior plane these were smaller than the posterior ones, except for Pelvic Tilt of 0.586° (p>012), falling within the limits of normality. The largest difference was found in Shoulder Alignment and Axillae Alignment of 0.655°. Regarding trunk asymmetry, it was found that the anterior values are higher than the posterior ones. Analyzing the posture for Body Alignment, most intended point of the trunk and Fontal Knee Angle, the correlations were weak ICC<0.4 and for Shoulder Alignment, Axillae Alignment, Pelvic Tilt, the correlations were good, ICC between 0.6 and 0.7. The ICC values for trunk asymmetry reflect a low correlation. Posture and body asymmetry assessment should be a periodic concern to allow timely corrections through the implementation and practice of effective and specialized prophylactic programs.