Skin lesion surface area and erythema intensity calculation for PASI determination using stereo vision

There are several tools to determine the severity of psoriasis and to track the effect of treatment, one of which is Psoriasis Area and Severity Index (PASI). PASI score calculation needs an accurate lesion area to total skin area ratio. To accurately determine these areas, 3D reconstruction from stereo vision has been used. From stereo pictures disparity maps and consecutively point clouds of only the object of interest are made, which for testing this method has been the arm. Transforming the colour space from RGB to YCbCr and inspecting the point clouds of each separate component, provides detection of red and lighter lesions, of which a point cloud can then be made. Using Delaunay triangulation on the point clouds results in a mesh of which the area can be calculated.

Tests have been conducted on light and dark skin shades, with red and pink lesions. It has also been tested for multiple lesions placed a few centimetres apart. The test results show that the formulated method works on both light and dark shade skin for red lesions, and detection of multiple lesions can be done as well. The method does not perform as well on lighter, pink lesions. The calculated lesion areas differed to the actual areas with an average factor of +37.45%