Hyperspectral Imaging for Noninvasive Diagnosis and Classification of Wounds, Burns, and Skin Lesions Tracking

EOS proposes to develop a hyperspectral imaging system (from the near UV through the near IR), for noninvasive acquisition and analysis of digital images of wounds, burns, and skin lesions. Furthermore, EOS proposes to assess the applicability of laser devices developed by Phillips Laboratory for elevation measurements of wounds to burns with structured illumination. EOS proposes this system as a new tool for: (1) quantitative monitoring of changes in lesions, burns, or wounds over time (e.g. in area, elevation, color, etc.); (2) reliable screening of the general population for skin cancers, including melanoma; (3) early burn diagnosis; and (4) wound healing assessment. The proposed effort will build on EOS's current efforts in development of sensor hardware and digital image processing for classification of skin lesions and will extend it to wound healing assessment and automatic burn classification. During Phase I, EOS will develop a hyperspectral sensor for imaging wounds and burns, select the optimal wavelength for elevation measurements, and design a commercial multiresolution system for lesion classification. During Phase II, two prototypes will be built and delivered to our medical collaborators who will acquire clinical image data on wounds and burns. EOS will use these data for development and testing of algorithms.