DERMAGUARD: Novel Wearable For Cutaneous Monitoring With Bio-printable Solutions

Video.

Skin damage is often invisible in its early stages and difficult to track as it heals. This gap between detection and treatment can lead to slower recovery and less effective care1. I wanted to explore how technology could close that gap. I was inspired by the limitation of current systems, which focus on identifying problems but not continuously monitoring them or improving treatment over time, current solutions are also invasive and are likely uncomfortable or painful. This led me to ask: What if detection, monitoring, and treatment could all be connected into one continuous non-invasive process?

To address this, I developed DERMAGUARD: a novel wearable for cutaneous monitoring with bio-printable solutions. The system captures images of skin and converts them into heatmap-style visualizations to highlight areas of damage. Image-based analysis is commonly used in medical diagnostics to enhance visibility of abnormalities 2. Unlike traditional approaches, it continues to monitor these areas over days or weeks, tracking how the skin changes. (NOTE: this model was mainly trained to identify and track melanoma and non melanoma skin cancer; this can always be expanded)

This evolving data can then be used to guide bio-printable skin model designing for better specific treatment, a growing field that enables the creation of customized tissue for regenerative purposes 3.

DERMAGUARD could benefit individuals with skin injuries, chronic conditions, or those requiring tissue repair, as well as researchers in biomedical and regenerative medicine.

By linking detection, continuous monitoring, and adaptive treatment, this project aims high.

To develop DERMAGUARD, I began by researching skin damage detection, image-based diagnostics, and regenerative medicine using trusted scientific sources to better understand how technology can support early identification and treatment4. I had also learned more about UV exposer and its effect on the skin layers.

After building this foundation, I designed an image capture system that could capture consistent images of the skin in a simple, non-invasive way. I then developed a method to convert these images into heatmap-style visualizations, allowing damaged or abnormal areas to stand out more clearly. Image-based analysis is widely used in medical diagnostics to improve the visibility of subtle abnormalities. I than proceeded to making a 3D modeling system that created 3D models from the images and UV information that can be used to show changes or as a regenerative skin treatment.

Once the concept was established, I focused on testing how the system could work over time. I Used big skin data sets like the International Skin Imaging Collaboration (ISIC) data set that contains many hundreds of diseased skin photos, to train/test my mode5. After the main testing part, I designed a wearable armband that was easy to wear everyday with an easy placement for wires to be mounted on.

The final setup included an imaging device (camera), a fixed positioning method to ensure consistency (wearable), and the correct boards, wires, and code to promise best results.

Results & Analysis

The DERMAGUARD prototype successfully demonstrated a complete system that turns skin images into visual and 3D models to represent surface changes. The system is designed as a wearable armband that captures images of the skin and processes them. It worked with real dataset images, and it consistently generated heatmaps and 3D models without consistant errors.

How the Prototype Works

The armband system works in three main steps:

Image Capture – The armband is designed to capture images of the skin over time. (Figure 5)

Heatmap Creation – The program compares brightness to the average skin to find differences and creates a intensity heatmap. (Figure 6)

3D Model Creation – The heatmap is turned into a 3D surface where each pixel becomes part of the shape, creating a potential regenerative medical solution to the effected epidermis. This part is then outsourced if needed. (Figure 7)

The Oled Screen than shows the the UV Risk score, and advises to change location according to UV score.

Analysis of Results

The results show that a simple method based on brightness differences can highlight surface changes effectively. While the system does not diagnose medical conditions with names, it demonstrates how a wearable device could track and regenerate the skin over time (Looking at differences only not identification of specific disease). When testing the model, the system from image capture to Oled Display was always working with no problem. The heatmap creation was where the model would sometimes make mistakes, the models tended to look at some "normal" parts of the skin and label them as diseased this would happen 6% of the time (Figure 8), which would than effect the regeneration of the epidermis.

Discussion & Conclusion

The results of this project are important because they show that a wearable system can convert simple skin images into clear visual and 3D regenerations of surface variation. Unlike many existing approaches that only display images, this project transforms visual data into structured outputs, making it easier to observe and find solutions over time. The wearable also proposes a much more cheaper solution, costing around $200, where as the current solutions can go up to $200,0006,7.

From the results, it can be concluded that a simple intensity-based method is effective for highlighting differences in skin surface patterns. The system consistently generated heatmaps and 3D models across multiple tests, showing that the pipeline is reliable and repeatable. This suggests that image-based monitoring can be a practical way to track changes without requiring complex or expensive equipment.

One key learning from this project is that even basic image processing techniques can produce meaningful and useful visualizations. By mapping pixel intensity to both color (heatmaps) and height (3D models), the system creates outputs that are easier to interpret than raw images alone.

These findings are important because they demonstrate how a wearable armband could support long-term monitoring by making small changes more visible over time. This could be especially useful in situations where gradual changes are difficult to notice8.

Overall, this project shows the potential of combining wearable technology with simple computational methods to create accessible and effective Regenerative tools, while providing a strong foundation for more advanced analysis.

This project could be extended by improving the wearable design from a fixed armband to a more flexible device that can be used on different areas of the body, allowing wider and more practical monitoring. The system could also capture images automatically over time for continuous tracking. Image processing could be improved using more advanced algorithms or machine learning, or a better camera could be used to improve detection. The 3D model could be made more detailed and realistic. Additionally, testing on a larger and more diverse set of images would improve reliability and accuracy.

I want to start by thanking my mentor, Usman Aziz, for helping me a lot with the technical side of this project, especially when I was stuck on the hardware and coding parts.

I also want to thank my mom, dad, and sisters for supporting with all the long nights and supporting me the whole time, even when things weren’t working. I really appreciate my grandparents, aunts, and uncles as well for always encouraging my ideas.

Lastly, thank you to everyone who helped in any way, whether it was giving feedback or just supporting the project. It really made a difference and helped me get to the final result.

References

Huang, Y., Xu, Y.-Q., Feng, S.-Y., Zhang, X., & Ni, J.-D. (2020a). LncRNA TDRG1 Promotes Proliferation, Invasion and Epithelial-Mesenchymal Transformation of Osteosarcoma Through PI3K/AKT Signal Pathway. Cancer Management and Research, Volume 12, 4531–4540. https://doi.org/10.2147/cmar.s248964

Huang, Y., Xu, Y.-Q., Feng, S.-Y., Zhang, X., & Ni, J.-D. (2020b). LncRNA TDRG1 Promotes Proliferation, Invasion and Epithelial-Mesenchymal Transformation of Osteosarcoma Through PI3K/AKT Signal Pathway. Cancer Management and Research, Volume 12, 4531–4540. https://doi.org/10.2147/cmar.s248964

Institute for Health Metrics and Evaluation. (2021a). GBD Compare. VizHub; Institute for Health Metrics and Evaluation. https://vizhub.healthdata.org/gbd-compare/

Huang, Y., Xu, Y.-Q., Feng, S.-Y., Zhang, X., & Ni, J.-D. (2020a). LncRNA TDRG1 Promotes Proliferation, Invasion and Epithelial-Mesenchymal Transformation of Osteosarcoma Through PI3K/AKT Signal Pathway. Cancer Management and Research, Volume 12, 4531–4540. https://doi.org/10.2147/cmar.s248964

ISIC Archive. (n.d.). Gallery.isic-Archive.com. https://gallery.isic-archive.com/#

Langreth, R., & Furlong, A. (2025, October 23). Micro-Doses of $200,000 Cancer Drugs Help Patients for Less. Bloomberg.com; Bloomberg. https://www.bloomberg.com/news/newsletters/2025-10-23/micro-dose-of-merck-s-keytruda-slashes-200-000-price-while-fighting-cancer

Karimkhani, C., Dellavalle, R. P., Coffeng, L. E., Flohr, C., Hay, R. J., Langan, S. M., Nsoesie, E. O., Ferrari, A. J., Erskine, H. E., Silverberg, J. I., Vos, T., & Naghavi, M. (2017). Global Skin Disease Morbidity and Mortality. JAMA Dermatology, 153(5), 406. https://doi.org/10.1001/jamadermatol.2016.5538

Rehm, R. G., Baum, H. R., & Barnett, P. D. (1982). Buoyant Convection Computed in a Vorticity, Stream-Function formulation. Journal of Research of the National Bureau of Standards, 87(2), 165. https://doi.org/10.6028/jres.087.013

Institute for Health Metrics and Evaluation. (2021b). Global Health Data Exchange (GHDx). Institute for Health Metrics and Evaluation; University of Washington. https://vizhub.healthdata.org/gbd-results/

VIDEOS

MostlyBuilds. (2023, August 10). 3D Modeling with Code! The best demo (OpenSCAD). YouTube. https://www.youtube.com/watch?v=KrFttd5D1cw

Paul Kry. (2014, September 10). Multi-layer skin simulation with adaptive constraints. YouTube. https://www.youtube.com/watch?v=HONEXF4jjZY