A method for kinematic analysis of chest wall motion is presented, based on a television-image processor that allows a three-dimensional assessment of volume change of the trunk by automatically computing the coordinates of several passive markers placed on relevant landmarks of the thorax and abdomen. The parallel computation used for the image processing allows for a real time recognition of the passive markers with the necessary accuracy. A geometric model also allows the online computation of the contribution to the chest volume by the different parts. For this purpose, the model presented here is based on 54 tetrahedrons that can be grouped into 9 compartments and 3 sections representing 1) upper thorax (mainly reflecting the action of neck and parasternal muscles and the effect of pleural pressure), 2) lower thorax (mainly reflecting the action of diaphragm and the effect of pleural and abdominal pressure), and 3) abdomen (mainly reflecting the actions of diaphragm and abdominal muscles). By this model, the volume can also be split into three vertical sections pointing out asymmetries between the right and left sides. The method is noninvasive, nonionizing, and leaves the subject maximum freedom of movement during the test, thus being suitable for routine clinical analysis. The monitoring of the subject can be prolonged in time and can be performed in different postures: standing, sitting, and supine. The method was tested on 12 healthy subjects showing its good accuracy, reliability, and reproducibility.
Real-time chest-wall-motion tracking by a single optical fibre grating: a prospective method for ventilator triggering
