Starting Dermi: From Research Projects to a Clinical Platform

Dermi Inc. was founded in Toronto, Canada, after years of academic research in dermatological imaging made one gap plain: the tools the research produced had no clinical system around them.

The Research That Preceded the Company

Before Dermi existed as a company, its founding work was a series of independent research projects in dermatological image analysis. Over several years, tools were developed for segmentation, classification, and image registration. Each addressed a real clinical need; the segmentation work was validated in peer-reviewed publications.

Segmentation models for automated body surface area estimation were validated against dermatologist manual assessment and published in the Journal of the American Academy of Dermatology in 2022. A multi-stage AI architecture combining segmentation, classification, and metric computation was introduced as the SPREAD framework in 2024. That same year, the AI PASS comparative study demonstrated AI-based psoriasis body surface area estimation with significantly lower error rates than three dermatologists. Dermi Atlas is a clinical documentation platform; these research methods are not part of it.

Each of these projects produced functional tools, but each existed in isolation, on its own infrastructure.

The Infrastructure Problem

The missing infrastructure was substantial: patient management to associate images with clinical records, structured image organization to support longitudinal comparison, audit trails for regulatory compliance, consent documentation integrated into the clinical workflow, and a deployment strategy that did not require dedicated IT staff.

EMR Integration

EMR integration was explored as an alternative to building a standalone platform. The reasoning was straightforward: if clinical infrastructure already existed in electronic medical record systems, the research tools could potentially be added as extensions or plugins rather than requiring a new platform.

This approach was ultimately abandoned. The dominant EMR platforms are closed-source systems, and their imaging capabilities are rudimentary, typically limited to basic photo attachment rather than structured clinical photography workflows. Building within these constraints would have required compromising the tools to fit the limitations of the host system.

The Decision to Build

The conclusion was that a dedicated platform was needed: a complete clinical photography system that practices could use as their primary imaging tool, rather than a plugin for an existing system. The platform is designed to operate alongside existing EMR systems as a specialized imaging tool, not to replace practice management or electronic health record software.

Architecture Decisions That Followed

A local-first, on-premises architecture was chosen to ensure data sovereignty. Patient imaging data would remain on computers the practice controls. This decision aligned with both regulatory requirements and the practical reality that many practices prefer direct control over sensitive patient data.

On-premises deployment introduced its own challenges. The platform needed to be accessible to practices without dedicated IT staff, which led to the development of Dermi Atlas Manager, a desktop application that handles the full deployment, certificate management for encrypted connections, backups, and updates. The goal was to make on-premises infrastructure as accessible as cloud-based alternatives without sacrificing data sovereignty.

Introducing Dermi Atlas: Clinical Photography Under Your Control covers the platform's capabilities and deployment model. The Atlas Cloud Demo provides free access to explore the platform's core features without installation.