Demonstrating the Capabilities of Hyperspectral Imaging for Dermatological Diagnostics

Researchers from the University of Greenwich have demonstrated the ability to track biological changes in a skin surrogate model using the Living Optics Hyperspectral Imaging Camera.

Identifying New Dermatological Diagnostic Techniques

Detecting changes in human skin enables the diagnosis of dermatological conditions, including early skin cancer detection. While traditional techniques such as dermoscopy and histopathology provide accurate data, they are often subjective and time-intensive. New dermatological diagnostic techniques aim to remove the need for experienced healthcare professionals to make subjective decisions while also reducing examination time.

A non-invasive imaging technique that measures reflectance differences across the visible and NIR wavelength range, Hyperspectral Imaging (HSI) is a leading candidate for monitoring biochemical changes in skin. HSI can reveal spectral signatures related to tissue composition, oxidative stress, and pigmentation anomalies, particularly relevant for skin cancer diagnosis.

Developing a Skin Surrogate Model

To prove these capabilities, researchers took a series of hyperspectral datasets using banana peel as a surrogate for human skin imaging. Both banana peel and human skin contain the enzyme tyrosinase, which regulates oxidative browning in bananas and plays a role in human melanin production.

The process of oxidative browning as bananas ripen has notable biochemical connections to skin pigmentation changes. Leveraging this connection allows banana peel to act as a cheap and ethically viable skin surrogate for demonstrating the proof-of-concept of HSI as a potential dermatological diagnostic technique.

Banana Peel Pigmentation Changes as Seen by the Living Optics Camera

Using the Living Optics Camera, hyperspectral images were collected over seven days, on days 1, 4, and 7. Spectral variations associated with pigmentation changes and structural transformations were observed across six regions of interest on three bananas.

Using the Living Optics SDK, the hyperspectral images were calibrated using a white and dark reference. Evaluation metrics were used to assess the quality of the reconstructed hyperspectral data in terms of spatial accuracy and spectral consistency.

Hyperspectral Imaging as a Dermatological Diagnostic Technique – A Proof-of-Concept

The surrogate model results prove that HSI can provide a non-invasive, cost-effective method of analysing pigmentation changes. The Living Optics Camera provided real-time hyperspectral datasets detailing how the biochemical makeup of the banana peel changed over time. The plot below shows the benefits of HSI over RGB and the level of spectral insight provided compared to a basic three-channel camera.

Spectral outputs from the Living Optics Camera detailed the pigmentation variation and bruising over 7 days in the banana peel. This analysis translates to dermatological conditions such as melasma, vitiligo, and melanoma. Tracking pigmentation over time could also help enable personalised dermatological treatments and disease monitoring.

Customer Insights

Professor Gianluca Tozzi from the University of Greenwich stated:

“By balancing high precision in capturing rich spectral data and flexibility of use, the Living Optics hyperspectral camera is ideal for exploring the unique spectral signature of biological tissues and supporting advanced AI tools for better healthcare.”

To learn more contact us today.