Skin conditions are among the most common medical concerns worldwide, yet access to a dermatologist can take weeks. AI-powered screening tools like DermaRash don't replace that visit — but they can help you understand what you're looking at and whether it warrants urgent attention.
How Image Classification Works
DermaRash uses a convolutional neural network (CNN) trained on labeled dermatological images. When you photograph a skin area, the model extracts visual features — texture, color distribution, edge patterns, lesion shape — and compares them against patterns learned during training.
The model outputs probability scores across a set of condition categories. A result like "82% consistent with contact dermatitis" means the visual features of your photo statistically resemble contact dermatitis cases in the training data — not that you have been diagnosed.
DermaRash is a screening tool, not a diagnostic device. All results include links to peer-reviewed sources (PubMed, AAD guidelines) so you can read the underlying research. Always consult a licensed dermatologist for diagnosis and treatment.
What the Model Can and Can't Do
- ✅ Flag patterns consistent with common conditions (eczema, psoriasis, rosacea, fungal infections)
- ✅ Distinguish between clearly different presentations (rash vs. dry skin vs. acne)
- ✅ Track changes in a skin area over time with photo history
- ❌ Diagnose melanoma or other skin cancers with clinical accuracy
- ❌ Account for symptoms you can't see (itch, pain, duration, triggers)
- ❌ Replace a dermatoscope exam or biopsy
Privacy and Data
Skin photos are processed on-device using Core ML. Nothing is sent to a server. Your photo history stays in your iCloud-encrypted local storage.
Cited Sources in Every Result
Every condition DermaRash identifies links directly to the American Academy of Dermatology (AAD) patient education pages and, where available, to PubMed abstracts of relevant clinical studies. This isn't common in consumer health apps — it's a deliberate design choice to keep you informed, not just reassured.
MobilApps.tech · iOS Developer: Evren Haznedaroglu