An Open-Source Robot That Creates 3D Maps of the Skin

Why OpenDerm Matters

More than 100,000 Americans are diagnosed with melanoma each year. This form of skin cancer has a five-year relative survival rate of nearly 100% when detected early, but that figure drops to about 34% once it reaches distant parts of the body. Roughly 70% of melanomas appear as entirely new lesions rather than changes in existing moles, which means the earliest visible sign may be a new spot only a few millimeters across. Detecting something that small across a person's entire skin surface requires a precise photographic record of how the skin looked before.

Total-body photography can provide that record, but existing systems remain expensive, limited in clinical availability, and often capture skin at resolutions too low to detect subtle changes. Marion Lepert, a PhD candidate in robotics at Stanford University advised by Jeannette Bohg, built OpenDerm as an open-source skin imaging robot to address those barriers. Open source means the complete hardware designs, software, CAD files, wiring schematics, bill of materials, and step-by-step build instructions are all publicly available for free, allowing any researcher, clinic, or institution to build, modify, or improve the system without licensing fees or restrictions. The total hardware cost is under $8,500, and the system captures skin at 78 pixels per millimeter, resolving details as fine as a mole's pigment network and individual surrounding hairs.

How the Open-Source Skin Imaging Robot Works

OpenDerm is a four-degree-of-freedom robotic gantry that moves a camera across the body with sub-millimeter positioning accuracy. Three linear axes position the sensor head within the workspace: one travels along the side rails, one moves across the top beam, and one sets the camera height. A fourth rotary axis tilts the sensor head to align the camera with the skin surface at each position.

The sensor head carries a Canon EOS R7 camera with a Canon RF 100mm macro lens, a Godox MF-R76 ring flash fitted with a cross-polarization filter that reduces specular glare, and two downward-facing laser distance sensors. The laser measurements allow the control system to maintain a consistent working distance and camera angle relative to the skin throughout the scan. At each imaging station, the robot adjusts its position using the laser data, then captures a high-resolution photograph. The complete pipeline aligns the overlapping images and reconstructs them as one continuous 3D surface. Registered scans taken at different times can then be compared point by point, revealing new lesions and identifying subtle changes in existing ones.

Exploring the Practical Benefits of Using OpenDerm

The system is built entirely from off-the-shelf parts. Unlike commercial total-body photography systems that rely on arrays of fixed cameras, OpenDerm uses a single camera moved by inexpensive actuators. This approach reduces hardware cost significantly while achieving higher spatial resolution than many wide-field systems. By moving close to the body and following its contours, the robot maintains consistent distance, angle, focus, and lighting across the entire skin surface.

Because the imaging task is entirely non-contact, the system avoids many of the manipulation challenges that make other robotics problems difficult. The robot does not need to grip, press, or physically interact with the patient. It simply positions the camera and captures images. The software is released under the MIT License, and the hardware design files are released under the CERN Open Hardware Licence Version 2. An interactive 3D scan viewer on the project website allows visitors to explore registered scans and inspect matched lesions across them.

How OpenDerm Compares to Other Approaches

Total-body photography currently uses three main approaches, each with different trade-offs. Wide-field systems photograph large areas of skin from a distance using fixed camera arrays. They are fast, capturing the body in seconds, but individual lesions may lack the resolution needed to detect subtle changes. Commercial examples include systems from Neko, Canfield VECTRA, DermSpectra, and FotoFinder.

Close-range robotic scanners like OpenDerm take a different approach. A high-resolution camera moves close to the skin and systematically across the body, maintaining controlled distance and viewing angle. This captures much finer detail than wide-field systems but requires more time because the body must be scanned sequentially. The third approach, guided smartphone imaging through apps like SkinIO, MoleMap, and Miiskin, requires the least hardware and can be used almost anywhere. However, lighting, distance, pose, and focus can vary between scans, making subtle longitudinal changes more difficult to measure reliably. OpenDerm sits in the middle category, offering dermoscopic-level detail at a fraction of the cost of commercial robotic systems.

The Three Gaps in Melanoma Early Detection

Improving melanoma early detection through total-body imaging requires progress in three areas. The first is higher resolution. Many existing systems use wide fields of view that sacrifice fine detail. Detecting subtle changes in a lesion's boundary, color, or internal structure requires imaging at a resolution that most wide-field systems do not provide. The second is longitudinal data. Most AI models for skin cancer detection are trained on isolated images of lesions already identified as suspicious. Training models to recognize earlier signs of melanoma will require repeated, high-resolution images of the same skin over time, which is exactly the type of dataset that a robotic scanner like OpenDerm is designed to produce. The third is accessibility. Total-body photography must become more affordable and widely available so that high-risk patients can receive frequent, standardized scans.

OpenDerm is currently designated as a research tool and is not classified as a medical device. Its value at this stage lies in demonstrating that high-resolution, reproducible skin imaging is achievable with open-source hardware at a cost point that could make longitudinal skin monitoring more accessible than it has been with existing commercial systems.