Duke University Engineers Develop AI-Powered Optical Imaging System to Monitor Deep Tissue Wound Healing

Duke University and Nokia Bell Labs develop an AI-powered optical system to non-invasively monitor tissue regeneration and blood flow in healing wounds.

By: AXL Media

Published: Mar 23, 2026, 5:34 AM EDT

Source: Information for this report was sourced from Duke University

Advancing Beyond Visual Inspection in Wound Care

Biomedical engineers have introduced a sophisticated alternative to the traditional, and often imprecise, method of visually inspecting skin wounds. While clinicians typically rely on surface-level measurements to track healing, what occurs beneath the skin often tells a different story of recovery or complication. By repurposing optical coherence tomography—a technology primarily used in ophthalmology—researchers at Duke University have created a way to visualize the architecture of healing tissue in three dimensions. This shift from subjective observation to data-driven analysis promises to eliminate the need for invasive biopsies that often reset the healing process by disrupting the wound site.

Synergy Between Nokia Bell Labs and Biomedical Engineering

The development of this platform was made possible through a multi-year partnership with Nokia Bell Labs, which provided the custom-built hardware and the computational framework necessary to interpret complex biological data. Professor Sharon Gerecht, the project lead, emphasized that the collaboration allowed her lab to merge their expertise in hydrogel therapies with advanced optical engineering. The resulting system uses light to penetrate the skin, capturing rich datasets of tissue structure that would be impossible to parse manually. This partnership highlights the increasing role of telecommunications and signal processing expertise in solving fundamental challenges in regenerative medicine.

Artificial Intelligence as a Quantitative Diagnostic Tool

The integration of artificial intelligence is the critical component that transforms raw imaging scans into actionable clinical insights. Trained on extensive datasets from the Gerecht lab, the AI models are designed to recognize and quantify specific markers of recovery, such as vascular density and tissue maturation. This automation removes human bias from the assessment, providing an objective score of how well a wound is closing beneath the surface. According to postdoctoral researcher Jiyeon Song, the AI enables real-time tracking of structural and vascular changes, offering a level of precision that manual image analysis cannot replicate.