Three-Dimensional Mesh Recovery from Common 2-Dimensional Pictures for Automated Assessment of Body Posture in Camptocormia

Abstract

Background: Three-dimensional (3D) human body estimation from common photographs is an evolving method in the field of computer vision. It has not yet been evaluated on postural disorders. We generated 3D models from 2-dimensional pictures of camptocormia patients to measure the bending angle of the trunk according to recommendations in the literature.

Methods: We used the Part Attention Regressor algorithm to generate 3D models from photographs of camptocormia patients' posture and validated the resulting angles against the gold standard. A total of 2 virtual human models with camptocormia were generated to evaluate the performance depending on the camera angle.

Results: The bending angle assessment using the 3D mesh correlated highly with the gold standard (R = 0.97, P < 0.05) and is robust to deviations of the camera angle.

Conclusions: The generation of 3D models offers a new method for assessing postural disorders. It is automated and robust to nonperfect pictures, and the result offers a comprehensive analysis beyond the bending angle.

Keywords: Parkinson; automated assessment; axial‐postural disorders; camptocormia; posture.

FIG. 1

(A) Example of a patient with Parkinson's disease with camptocormia. (B) The 3‐dimensional mesh is overlayed, and the green dots mark the automatically identified key points for the pose estimation algorithm used prior to mesh fitting. Furthermore, the points for manual pose estimation used with the NeuroPostureApp are illustrated in yellow. (C,D) The extracted mesh in the “Blender” software from a back and more frontal view for illustration. The algorithm used is not suitable to specifically estimate the hand position; therefore, these are not estimated correctly. LM, lateral malleolus.

FIG. 2

(A) Highly significant Pearson correlation between the angles of the trunk flexion measured in 40 pictures using the manual versus the automated mesh‐based method. (B) The associated Bland‐Altmann plot is shown. (C,D) Angle estimation on the virtual human posed in a moderate camptocormia posture (C) with a bending angle of 55 degrees measured by the manual method and an extreme camptocormia posture (D) with 77 degrees of forward flexion. The figure shows that the method does not provide reliable angle estimates if the image is created from an extreme front or rear perspective. The dark dotted lines indicate a range of 5 degrees above and below the manually measured value. CC, camptocormia.