SkinVision Files IPO Prospectus: AI-Powered App for Early Melanoma Detection via Smartphone Imaging

"The 'Deadly Threat' in Moles"

As summer arrives, beauty enthusiasts invariably layer on sunscreen or avoid sun exposure altogether. Yet, this does not prevent the emergence of those pesky dark spots or moles. While such blemishes may be less concerning when hidden beneath clothing, their appearance on the face can be truly distressing—after all, not everyone can boast a mole as iconic and attractive as Marilyn Monroe’s beauty mark.

In fact, dark spots that develop under intense sun exposure not only affect appearance but also pose many potential risks, as they may harbor cancerous cells. Melanocytic nevi (especially those exhibiting abnormal growth) are early signs of skin cancer, and malignant melanoma—the most lethal form of skin cancer—often develops from seemingly innocuous moles. If diagnosed at an early stage and treated promptly, up to 95% of skin cancer patients can be cured or achieve long-term survival; however, if detected only at a late stage, the long-term survival rate drops to just 15%. Caucasians naturally lack sufficient melanin to protect against ultraviolet (UV) radiation from sunlight, resulting in a higher incidence of skin cancer compared to Asians and Africans.

According to data from Cancer Research UK, 2,000 people in the United Kingdom alone died from this disease in 2012. In the UK, the annual cost of treating melanoma exceeds £10 billion; however, early detection can significantly reduce the treatment costs for skin cancer. To address this, the SkinVision team from the University of Bucharest in Romania launched an iOS and Android scanning application designed to assess whether users' moles pose potential risks. The app, named SkinVision, can detect 73% of melanomas using only images. In fact, mobile applications for self-examination of moles have substantial market potential in Europe and North America. It is reported that in 2013, SkinVision passed scientific validation trials at the Ludwig-Maximilians-Universität clinic in Munich, Germany, and is the first and only skin cancer diagnosis app to comply with the EU CE certification.

“Lethal Threat” to Lifeline

While this app cannot replace a dermatologist, it raises awareness of melanoma risks. The seemingly simple features of SkinVision are the result of hard-won efforts—a product of cross-disciplinary collaboration among multiple dermatology experts, mathematicians, and computer scientists. SkinVision’s original design intent—to facilitate condition monitoring and prompt timely medical consultation—is commendable. Furthermore, the application has been developed with rigor; by establishing personal profiles, it conveys a high degree of accuracy in its test results. Additionally, its underlying principle is worthy of recognition. SkinVision has built a database containing over 100,000 moles, 4,000 of which have been diagnosed or reviewed by dermatology experts. Mathematicians and computer scientists continue to refine SkinVision’s recognition algorithms based on this data. Employing “fractal analysis” technology, the algorithm identifies moles and surrounding skin tissue according to size, diameter, and numerous other parameters, while detecting whether the moles exhibit abnormal growth patterns.

Mircea Popa, Co-founder of Product, stated that because the skin is composed of a large number of cells and undergoes continuous renewal according to specific patterns, it can be measured using mathematical models. Harmful cells undergo changes, and their morphological characteristics during proliferation and replication differ from those of normal cells. The primary distinctions lie in color and shape, which are the two features detected by the app.

A team of lecturers from Ludwig Maximilian University of Munich described the application as “an effective tool for the predictive assessment of moles.” Their investigation revealed that, among all melanoma cases collected by the application to date, 73,100 had been correctly identified.

You might wonder how an app analyzes moles. In fact, the user-side operation is very convenient: simply take a clear photo of the mole with your smartphone (the SkinVision website provides detailed photographic guidance), and you will receive an analysis within seconds. After the initial assessment, users can utilize the archiving feature to track changes over time. In addition to mole analysis, the app offers preventive care and health information. Most notably, it displays the UV index for your area, helping you determine when to double up on SPF protection (although we recommend daily use of SPF 30+ regardless of weather conditions). Furthermore, the app features an easy-to-use sharing function that allows you to send your data to a doctor. With just one tap, the app displays a map marking all nearby dermatology clinics.

Let's take a look at the specific steps for implementation:

Open the app. You must first register on the website; once completed, you can log in directly without email verification.

After agreeing to various agreements, the app displayed an interface for creating a profile. Enter your first name, last name, age, and select your gender and country as required. Click “Done” to complete the profile setup.

For greater accuracy, users are advised to confirm UV intensity via geolocation and complete the skin type assessment. To do so, navigate to the “TOOLS” section, select “UV Exposure” and “Skin Type,” and answer eight brief questions regarding skin tone and sun exposure to obtain reference results.





It is evident that the UV intensity in Beijing is currently very high.

Start taking photos. Capture one or more moles, ensuring the images are as clear as possible. You can then drag and stretch to adjust the photo, making sure the user's mole remains within the frame. Click “Analyze” to upload it to the app’s server, and wait approximately 10 seconds for the upload and analysis to complete.





However, the software often displays an “Analysis error” prompt. This is typically due to issues with lighting or image clarity. Please adjust the settings or zoom in, then re-upload the image and try again.

Analysis complete; results displayed. The findings are presented using a specialized sci-fi visualization, which can be switched to the original image by tapping the screen. The app categorizes users’ moles into three risk levels—low, medium, and high—indicated by green, yellow, and red, respectively, and provides corresponding recommendations.

In fact, the analysis result shown in the image above was based on black dots drawn on the editor’s body with a black pen. Previously, when testing the editor’s actual moles, the analysis consistently returned “Analysis error.”

If users wish to consult a physician, the app also provides a selection of hospitals for them to choose from. By clicking the “Send” button in the next step of the diagnosis process, their information can be sent directly to the physician’s email address.

However, this application still has some issues that need to be addressed: it places considerable demands on practical operation, and its recognition algorithm is not yet sufficiently intelligent. In one instance, a user performed two separate scans of the same mole; due to variations in image quality, the test results fluctuated significantly, ranging from medium to low risk. Furthermore, the majority of test results were predominantly yellow or green, failing to provide red alerts when they were truly warranted.

How Can Vitality Become Business Opportunity?

In 2012, SkinVision secured investment from the German investment firm Personal Health Solutions. From its launch in 2012 to November 2014, the app had been downloaded by approximately 100,000 users, who contributed 175,000 photos to the product database for future development.

SkinVision is available for both iOS and Android. Currently, the app offers a free one-month trial, during which users can access all features. After the trial period, users can continue using the service through a subscription, with fees varying based on the subscription duration. Users can choose from one-month, three-month, or one-year subscription plans.

(Editor: Zheng Qi)