Smartphones and Skin Cancer Detection
For the past thirty years, the incidence rates for melanoma, the deadliest form of skin cancer, have continued to increase in the United States. Even though mid-century practices of tanning oils and foil mirrors have dwindled and skin cancer awareness campaigns have been stepped up, many millennials still flock to tanning salons in order to maintain a glow all year round. Sadly, the continued exposure to ultraviolet (UV) rays has led to higher risks of melanoma, especially in females under forty. This type of cancer, which develops in cells that produce melanin (the pigment that gives our skin its color) can develop anywhere on the body, although it is most commonly found on areas prone to sun exposure including the back, legs, arms, and face. One in five Americans will be diagnosed with skin cancer in their lifetime, and on average one American dies from skin cancer every hour.
A few years ago, Miad Faezipour, Ph.D., assistant professor of computer science and engineering and biomedical engineering, and Buket Barkana, Ph.D., associate professor of electrical engineering, joined forces with then Computer Science and Engineering doctoral student Omar Abuzaghleh to begin development of a novel, smartphone-based virtual reality system to aid in melanoma detection and prevention. Their application, SKINcure, focuses on the analysis of suspicious moles and lesions and on prevention notification. In a growing technological world, they understand that a convenient and user-friendly application is likely to become an important tool in the prevention and screening of this most deadly form of skin cancer.
Through the SKINcure application system, the user is able to analyze suspicious moles and lesions by snapping a smartphone photo and uploading to the application. For better results, the SKINcure app is aided by a handy-scope camera which easily attaches to the iPhone’s camera lens. This allows for a higher quality image to be captured and analyzed by the SKINcure application system. Then, the photo of the mole in question is compared to PH2 dermoscopic images in a comprehensive database. The database images have been obtained under the same 20x magnification conditions as the images captured through the SKINcure application. The image database contains 200 dermoscopic images of lesions, including 80 normal moles, 80 atypical moles, and 40 melanomas. The diversity of the images in the database allows for a better analysis of the images collected from the user. SKINcure then analyzes the mole using the dermoscopic image database and classifies it as either “normal,” “atypical,” or “melanoma.” If the mole is classified as atypical or melanoma, the user is notified to seek medical help immediately in order to increase the chances of successful treatment options.
In addition, the system is able to capture user environmental data, UV radiation level and skin images to conduct a risk assessment and alert the user in real-time to prevent risks associated with developing skin cancer. SKINcure detects sunlight from the user environment using the smart phone GPS system and handy-scope camera. By mapping the user location and the time of day, an accurate ultraviolet radiation level can be calculated. At the time of initial registration, the user will be asked to select a burn frequency that best describes his or her skin: “rarely,” “sometimes,” “usually,” or “always.” After the data is captured through the application it will alert the user in real-time when the user encounters exposure to high UV radiation and over sun exposure, allowing the user to reapply sunscreen, seek shade, and use other methods of prevention. The outcome of the research platform is intended to help users prevent skin cancer by triggering the real-time alert that informs users when to avoid exposure to harmful UV radiation and to help with early detection of melanoma in order to increase the chances of successful treatment options.
Faezipour joined the engineering faculty in July of 2011 and has quickly established a solid research focus through receiving a number of institutional Seed Money Grants. Her research interests lie in the broad area of biomedical signal processing and behavior analysis techniques, high-speed packet processing architectures, and digital/ embedded systems.
Barkana joined the engineering faculty in 2007 in the department of electrical engineering. This research projected was funded through a Seed Money Grant awarded by the Faculty Research Council. Her research interests include all aspects of signal processing including speech, non-speech, and bio-signals, biomedical image processing, and innovations in K-12 STEM education. She is the author of over sixty published journal and conference articles.
Abuzaghleh graduated with his Ph.D. in Computer Science and Engineering in May 2015 and is employed by the UB School of Engineering as adjunct faculty and assistant lab manager. In April 2016, Abuzaghleh was the recipient of a UB Faculty Research Day Faculty Award for his dissertation research to develop SKINcure, which has already received a provisional patent.
The research was funded in part by a Seed Money Grant from the University of Bridgeport.