Ideaz

The problem:

Melanoma is a type of skin cancer that is rapidly growing worldwide. Although it is usually not deadly, it still has many symptoms like unhealable sores, irritated skin, etc. Similar to other cancers, the easiest and best way to treat it is all dependent on early detection of the cancer. Despite the existence of some apps that analyze skin symptoms to see if melanoma is forming, many patients still discover the cancer far too late. Current solutions also have limitations like lacking clinical validation, failing to analyze many diverse skin tones, or being unable to provide proper follow up care. 

The solution:

This is why I propose an AI-powered image processing API that can analyze and monitor a patient's skin. Melanoma is detectable upon 5 rules, ABCDE of moles: 

• Asymmetry: One half doesn’t match the other half
• Border: irregular borders 
• Color: Colors changing across different moles
• Diameter: larger than 6mm
• Evolution: changing in size, shape or behavior

Patients can use their smartphone camera and upload it to the API and the AI, which is trained on a database of multiple varied skin tones and their behaviors, would analyze the photo with image processing and determine a risk factor of the patient. For patients experiencing a medium risk factor, we would suggest some lifestyle tips on how to move forward and continue tracking their progress. For patients experiencing a high risk factor, we would recommend dermatologists to visit nearby in hope that they can fix their condition. 

Who benefits:

Any patient experiencing skin irregularities.

Doctors, as it helps catch skin cancer earlier and prevents more work needed to be done to treat the patient

Hospitals, as I am selling this API to dermatologists to be able to easily treat their patients without the attention of a doctor all the time.

Why this matters to me:

Cancer is of course a horrible disease to be diagnosed with. Having the ability to be able to catch it earlier greatly increases the chance of it being cured.

Technical Details:

I would make an API that could be integrated with a hospital website or app. Many apps now are made to track the heart, lungs, brain etc, and it would be very simple to add one more feature that I could sell to them.