About the project

We're working on the terms and conditions, you can read them here:

    Dear colleagues, patients and beta testers!

    First of all, thank you for visiting the page.

    This online application is a result of collaboration in between two individuals trying to advance the technology, Artificial Intelligence, user friendly design into the medical field. Our objective is to simplify and improve the methods of early diagnostics of certain types of skin cancer for both patients and doctors, to spread the awareness of the illness and to improve health of the general population.

    Melanoma of the skin is a tumor with a predictable course, but with early diagnosis, the 5-year survival rate reaches almost 99%.

    Symptoms of the diagnosis of skin melanoma are known, but its diagnosis at an early stage is far from ideal. This is facilitated by many reasons: a rare occurrence in the practice of primary contact physicians in certain areas, the availability of qualified assistance.

    There are many publications that show that the diagnosis of skin melanoma using neural networks allows for a more accurate diagnosis than practicing dermatologists.

    We, as a group of researchers, are trying to use available artificial intelligence technologies for the benefit of the patients for the early diagnosis of skin melanoma. For this, we have collected all our experience and knowledge in diagnostics and have developed an autonomous expert system.

    The application uses online form to either upload a photo from your local database or to capture a photo from the web camera embedded into your device. 

    After capturing the photo or uploading the photo into the application, the image is then being uploaded to the algorithm to analyze and return it's prediction based on the gathered information.

    The current iteration of the algorithm has been trained on more than 1 thousand photographs captured through dermatoscope with immersion. The database of the images had been carefully sorted prior to training and only relevant images have been used to train the algorithm.

    Currently, the latest iteration of the model is being hosted in the Microsoft Azure ® cloud, along with the dedicated server that is capable of processing a large amount of transactions.

Our team:

- Neretin Evgeny Yurievich

Oncologist of the highest category of the consultative department No. 1 in the State institution of health care Samara Regional Clinical Oncological Dispensary, Associate Professor of the Department of Surgery.

- Yarik Sychov

Master of Engineering, currently working as a Chief Engineer in Maritime industry.

Current version of the application is in BETA testing mode. 

We're constantly working on improving the accuracy of the trained algorithm, getting more training data is essential. Please get in touch if you have access to data that can be used by us.

Thank you!

Last update information.

Finished training on 12/6/2020, 06:29:10 PM using General (compact) domain

Iteration id: ****61f1e74

Classification type: Multiclass (Single tag per image)

Latest application update: 31 January 2021