In vivo skin elastography with high-definition optical videos
Creator
Zhang Y; Brodell R T; Mostow E N; Vinyard C J; Marie H
Publisher
Skin Research and Technology
Date
2009
2009-08
Description
Background/aims Continuous measurements of biomechanical properties of skin provide potentially valuable information to dermatologists for both clinical diagnosis and quantitative assessment of therapy. This paper presents an experimental study on in vivo imaging of skin elastic properties using high-definition optical videos. The objective is to (i) investigate whether skin property abnormalities can be detected in the computed strain elastograms, (ii) quantify property abnormalities with a Relative Strain Index (RSI), so that an objective rating system can be established, (iii) determine whether certain skin diseases are more amenable to optical elastography and (iv) identify factors that may have an adverse impact on the quality of strain elastograms. Methods There are three steps in optical skin elastography: (i) skin deformations are recorded in a video sequence using a high-definition camcorder, (ii) a dense motion field between two adjacent video frames is obtained using a robust optical flow algorithm, with which a cumulative motion field between two frames of a larger interval is derived and (iii) a strain elastogram is computed by applying two weighted gradient filters to the cumulative motion data. Results Experiments were carried out using videos of 25 patients. In the three cases presented in this article (hypertrophic lichen planus, seborrheic keratosis and psoriasis vulgaris), abnormal tissues associated with the skin diseases were successfully identified in the elastograms. There exists a good correspondence between the shape of property abnormalities and the area of diseased skin. The computed RSI gives a quantitative measure of the magnitude of property abnormalities that is consistent with the skin stiffness observed on clinical examinations. Conclusions Optical elastography is a promising imaging modality that is capable of capturing disease-induced property changes. Its main advantage is that an elastogram presents a continuous description of the spatial variation of skin properties on the pixel level that would otherwise be impossible with other sensors. Its value will be further enhanced when used with a point-wise measuring device such as a cutometer that yields absolute elasticity values.