Anuradha Godavarty
Florida International University | USA
Title: Automatic segmentation of lower extremity ulcers using near-infrared optical imaging
Biography
Biography: Anuradha Godavarty
Abstract
Lower extremity ulcers are one of the most common complications in diabetes patients. Clinical studies have shown that reduction in wound size of < 40% within 4 weeks is an acceptable progress of the healing process. To date, visual inspection is usually adopted to monitor the wound healing process and wound size is measured from the surface of the images. A Near-InfraRed (NIR) optical imaging approach has been developed to internally inspect the wound by non-contact imaging to differentiate the healing of the non-healing wounds. Herein, an automated image segmentation method was used for the first time to quantitatively measure the wound size from NIR and white light images. NIR wound segmentation was carried out on multiple venous leg ulcers and diabetic foot ulcers (FIU-IRB approved study). Upon applying graph cuts segmentation algorithm, the wound shape demarcated from NIR images closely matches the white light images (but not completely). This revealed that the NIR images are possibly providing tissue oxygenation information from the internal of the wound and not the surface of the wound. Our ongoing efforts are to co-register the white light and NIR images for comparison of the wound demarcated regions and quantification of the wound area