SkinCrab

SkinCrab is an app that would allow the early detection of skin cancer  through a photograph using Artificial Intelligence. 

SkinCrab is a mobile application that uses Artificial Intelligence for early detection of skin cancer. This works through photographs that the users take with their smartphones.

Unlike existing apps, SkinCrab uses an algorithm capable of detecting multiple types of skin cancer and other lesions and provides a more personalized experience through quizzes as well as the ability to perform a general skin health self-exam.

The application is free and once the diagnosis is generated, it recommends the user a series of specialists according to the type of skin injury. In addition, it includes a guide in this way, it seeks to improve the accuracy of diagnoses and prevent the disease from reaching more advanced stages.

Seven Costa Ricans a day are diagnosed with skin cancer, around 3 million people are diagnosed per year worldwide and in 2019 only in Costa Rica there were more than two thousand of them. Skin cancer is a disease whose diagnosis involves a large investment in terms of time and resources, to which not everyone has access. 

How advanced this disease is, and what stage it is in at the time of diagnosis is decisive to define the development that the patient will have and its possibility of improvement, since in its last stages this condition extends to other parts of the body, such as the lungs, liver, brain, or bones.

Skin cancer is so dangerous, but if it is treated on time it can be completely treatable. However, many times patients present themselves to medical centers too late. That's why something that makes early detection accesible, such as SkinCrab, is super important. It is one of the biggest challenges of modern dermatology, for it means a risk to health and life.

We must act in time, which is the purpose of SkinCrab: to offer an alternative that allows people, especially those with low resources, to self-detect this disease as soon as possible and only needing a cell phone.

We would like to help the entire population in general, but especially those people who do not have the economic resources to go to a medical appointment every time they have a suspicion of a change in the skin or who have a family history of skin cancer.

Diagnosis processes for skin cancer have high prices that are not accessible to the entire population, and the free alternatives offered by some health centers can take a long time and in skin cancer, early diagnosis is extremely important. That is why SkinCrab would offer the population an option to be able to get ahead in the diagnosis process and thus prevent it from reaching critical stages, by allowing self-detection of this disease, requiring only a cell phone. Thus, anyone who owns a smartphone can access this resource. In addition, campaigns will be sought from time to time where technological devices are made available so that people with very low resources who do not have cell phones can also take advantage of this tool. In this way, the target population would be anyone who cares about the well-being of their skin and wants to use the tool to maintain constant control of this disease, regardless of their resources.

In the first place, a survey was created where we focused on the knowledge and opinion of the population about the technology and artificial intelligence involved in skin cancer and its consequences. From which we got very worrying answers. So we decided to do a deeper investigation about the current solution. And create statistics about the results of this solution. Of which we got poor results.

In this way we look for a way to combine current solutions and take into account the opinions of people about existing applications. So that those feedbacks would be more hungry to what the users need and create our own solution. In this way, inform the population and help them commit to their health.

We are currently in the prototyping phase. Where we already have a low-fidelity prototype and we are currently working hard on creating the application to make it 100% functional. Since for now we only have certain parts of the application such as registration, forms, the network of doctors, the creation of the diagnosis and others.
In addition to the Machine Learning model where the images are classified into three categories (the most common types of cancer): melanoma, epidermoid carcinoma and basal cell carcinoma.

We hope in a few months to be able to finish the prototyping phase and be able to test it and have the validation of both doctors and the population.