AI researcher develops low-cost skin cancer detection tool for remote areas

A researcher at Heriot-Watt University is leveraging artificial intelligence to address critical gaps in medical care for patients in remote regions, offering a potential breakthrough in early skin cancer detection. Tess Watt, a Ph.D. candidate in the School of Mathematical and Computer Sciences in Edinburgh, has spent the past two years developing AI tools designed to diagnose skin conditions—including skin cancer—where access to dermatologists is limited. Her goal is to enable individuals to monitor their health from home, reducing delays in diagnosis that often arise due to long GP wait times. The project, part of a collaborative effort with academics from London South Bank University, Edinburgh Napier University, and the Foundation for Research and Technology—Hellas in Greece, is the first of its kind to integrate AI-driven medical diagnosis with a focus on remote communities. A prototype of her system has been showcased at Heriot-Watt’s advanced health and care technologies suite, demonstrating how AI can analyze images of skin lesions and identify potential risks for further medical evaluation. What distinguishes the system is its functionality offline. It operates using low-cost Raspberry Pi devices, which act as portable, energy-efficient computers capable of storing large datasets. Patients would use a small camera connected to the device to capture images of skin issues, which the AI then processes in real time by comparing them to thousands of medical images. Results are shared with local general practitioners to initiate treatment. The tool currently achieves 85% diagnostic accuracy, with plans to enhance this by incorporating more skin lesion datasets and advanced machine learning models. Though the technology has not yet been tested in clinical settings, Watt is engaging with NHS Scotland to secure ethical approval for a pilot project, which she hopes to launch within the next one to two years. She acknowledges that medical innovations often take years to transition from prototypes to real-world applications. The long-term aim is to deploy the system first in remote parts of Scotland before expanding globally to regions with scarce dermatological resources. It could also assist patients who are unable to travel, allowing family members to help capture and submit images for review. Watt’s work in health AI stems from her earlier research in accessible translation technologies, combined with her studies in Tiny Machine Learning and exploration of underdeveloped AI applications in medicine. “There’s a lot of focus on AI for X-rays and MRIs, but fewer efforts on skin photographs. I saw an opportunity there,” she noted. Dr. Christos Chrysoulas, Watt’s academic supervisor and an associate professor of computer science, emphasized the importance of designing resilient systems for healthcare. “E-health devices must function independently of external connectivity to ensure patient safety and service continuity,” he said. He highlighted the need for core diagnostic and therapeutic capabilities to remain operational even during network outages, adhering to strict safety and regulatory standards. This focus on reliability is critical for resource-limited areas with unreliable internet access. Watt’s research aligns with Heriot-Watt University’s Global Research Institute (GRI) in Health and Care Technologies, which prioritizes solutions for pressing global health challenges. The institute, active across campuses in Scotland, Dubai, and Malaysia, is also advancing initiatives like “One Health,” a research theme addressing infectious diseases, antimicrobial resistance, and epidemics through interdisciplinary approaches. As the project progresses, Watt aims to ensure her work translates into practical healthcare improvements by the time she completes her Ph.D. in three years. Her system represents a step toward equitable, efficient care, addressing U.K. health service backlogs while offering scalable solutions for underserved populations worldwide. The collaboration between academia and industry partners underscores the growing role of AI in transforming medical accessibility, particularly in regions where traditional infrastructure is lacking.