DermScan
Deep Learning Skin Cancer Detection & DroidCam Mobile App
Medical-grade mobile application leveraging Deep Convolutional Neural Networks (CNN) and DroidCam camera streams for early skin lesion analysis and dermatological screening.
- Year
- 2026
- Duration
- 6 months
- Team
- 6 people
- Sector
- Healthcare
0.0%
Detection accuracy
Deep CNN model sensitivity score
0.0s
Scan time
Instant lesion classification
0K+
Scans completed
Screening scans performed
0%
Dermatologist sign-off
Diagnostic concurrence rate
DermScan Health
HealthTech AI company developing non-invasive screening tools.
Enabling early skin cancer detection using accessible AI technology.
Logo concept
Shield icon enclosing a digital dermal scan line.
#2563EB#3B82F6Founded
2024
Headquarters
India
Employees
15
Sector
Healthcare
Every surface we shipped
11 designed screens across 2 deliverables, rendered live rather than captured as static images.
DermScan Home
Camera scanner, risk history, and dermatologist locator.
Lesion Analysis
Deep CNN skin risk classification.
Model Accuracy
Sensitivity & specificity metrics.
Dermatologist Queue
Expert review & consultation booking.
Scan History
Tracked lesion changes over time.
Medical Summary
Exportable PDF report for doctors.
Settings
Organization, team, and integration configuration.
Mobile application
Sign In
Create Account
Notifications
Profile
Across every form factor
The same design system, rendered at each breakpoint it has to survive.
The challenge
Early skin cancer screening is often delayed due to lack of accessible dermatological tools.
Research
- Dataset curation from ISIC international skin imaging collaboration
The solution
Developed DermScan, an AI-powered mobile scanner using Deep CNNs and DroidCam integration.
UX decisions
Color-coded risk indicators
Clear guidance for users to seek immediate medical consultation when needed.
Features
- Deep CNN lesion classifier
- DroidCam camera integration
- Dermatologist report PDF export
- Scan history timeline
Technology
Architecture
TensorFlow Deep CNN model deployed via Dockerized Python FastAPI with React Native mobile client.
Results
Achieved 96.8% detection sensitivity across 25,000+ completed screening scans.
Lessons learned
- 01Providing clear disclaimer and doctor referral workflows builds essential clinical trust.
How the engagement ran
6 months across 3 phases with a team of 6.
CNN Model Training
10 weeksTrained Deep CNN on ISIC skin cancer dataset.
DermScan CNN ModelApp & DroidCam Sync
8 weeksIntegrated high-res camera feed and DroidCam streaming.
DermScan AppClinical Validation
6 weeksValidated model performance with dermatologist partners.
Clinical Audit Report
Have a problem shaped like this one?
We start every engagement with a paid discovery sprint. You get an architecture assessment and a delivery plan — whether or not you continue with us.
Related engagements
About this case study: DermScan Health is a fictional client. This engagement, its metrics, and its quotes are illustrative work product created to demonstrate our delivery approach, architecture reasoning, and design process. They do not describe a real customer.