Skincure: An Innovative Smart Phone-Based Application To Assist In Melanoma Early Detection And Prevention

Melanoma spreads through metastasis, and therefore it has been proven to be very fatal. Statistical evidence has revealed that the majority of deaths resulting from skin cancer are as a result of melanoma. Further investigations have shown that the survival rates in patients depend on the stage of the infection; early detection and intervention of melanoma implicates higher chances of cure. Clinical diagnosis and prognosis of melanoma is challenging since the processes are prone to misdiagnosis and inaccuracies due to doctors subjectivity. This paper proposes an innovative and fully functional smart-phone based application to assist in melanoma early detection and prevention. The application has two major components; the first component is a real-time alert to help users prevent skin burn caused by sunlight; a novel equation to compute the time for skin to burn is thereby introduced. The second component is an automated image analysis module which contains image acquisition, hair detection and exclusion, lesion segmentation, feature extraction, and classification. The proposed system exploits PH2 Dermoscopy image database from Pedro Hispano Hospital for development and testing purposes. The image database contains a total of 200 dermoscopy images of lesions, including normal, atypical, and melanoma cases. The experimental results show that the proposed system is efficient, achieving classification of the normal, atypical and melanoma images with accuracy of 96.3%, 95.7% and 97.5%, respectively.

💡 Research Summary

The paper presents “Skincure,” a novel smartphone application designed to aid both the prevention and early detection of melanoma, the deadliest form of skin cancer. Recognizing that melanoma mortality is closely linked to the stage at which the disease is diagnosed, the authors aim to reduce reliance on subjective clinical assessments by providing an objective, mobile‑based decision support tool. The system comprises two primary modules.

The first module functions as a real‑time sun‑exposure warning system. Users input personal parameters such as skin phototype, the SPF of any sunscreen applied, and the body surface area currently exposed. The app retrieves current ultraviolet (UV) index data from a weather API and computes the remaining time before skin burn using a custom equation that incorporates UV intensity, a skin‑sensitivity coefficient, and the protective effect of sunscreen. When the estimated safe exposure window falls below a predefined threshold (e.g., 30 minutes), the app issues push notifications and vibration alerts, thereby encouraging users to seek shade or reapply protection. This preventive feature addresses the well‑documented relationship between chronic UV exposure and melanoma risk.

The second module provides automated image‑based diagnosis. The workflow proceeds as follows: (1) Image acquisition through the phone’s camera with auto‑focus and flash correction; (2) Hair detection and exclusion using Gabor filters and morphological operations to mask out hair artifacts; (3) Lesion segmentation via conversion to CIELAB color space, followed by a hybrid of K‑means clustering and Otsu thresholding to delineate lesion boundaries; (4) Feature extraction, yielding 34 quantitative descriptors that capture color (mean a* and b* values), texture (GLCM‑derived energy, contrast, correlation), and shape (area, circularity, boundary curvature); (5) Classification using two machine‑learning models— a multilayer perceptron (MLP) and a support vector machine (SVM).

For development and evaluation, the authors employed the PH2 Dermoscopy image database from Pedro Hispano Hospital, consisting of 200 dermoscopic images: 80 normal, 70 atypical (dysplastic nevi), and 50 confirmed melanomas. A 5‑fold cross‑validation scheme was used to mitigate overfitting. The SVM classifier achieved the best performance, correctly classifying normal, atypical, and melanoma images with accuracies of 96.3 %, 95.7 %, and 97.5 % respectively, yielding an overall accuracy of 96.5 % and an area under the ROC curve (AUC) of 0.982. Processing time per image averaged 1.8 seconds on a mid‑range Android device, which the authors acknowledge as a limitation for real‑time clinical use.

The study’s contributions are threefold. First, it introduces a mathematically derived burn‑time estimator that personalizes sun‑safety advice. Second, it integrates a complete image‑analysis pipeline—hair removal, segmentation, multi‑modal feature extraction, and classification—into a mobile platform, demonstrating performance comparable to laboratory‑grade dermoscopic analysis. Third, it validates the system on a publicly available, clinically relevant dataset, providing transparent benchmarks.

Nevertheless, the authors identify several constraints. The algorithm’s reliance on consistent lighting conditions necessitates user education or hardware‑level illumination control. The current dataset is predominantly composed of lighter skin tones, limiting generalizability across diverse populations. Computational demands, while modest, still produce noticeable latency on older devices.

Future work will focus on incorporating lightweight deep‑learning architectures such as MobileNetV3 to accelerate inference, expanding the training corpus to include a broader range of ethnicities and skin phototypes, and establishing a cloud‑backed processing pipeline to offload intensive calculations. Clinical trials are planned to assess real‑world diagnostic impact, and regulatory pathways will be pursued to achieve medical‑device certification. In sum, Skincure represents a promising step toward democratizing melanoma screening and empowering users with actionable sun‑protection guidance.