If you are a researcher, you can collect large amounts of data when studying a specific subject. However, a health-related startup called Seed is conducting a project to create a stool image database by collecting 100,000 stool images.
In the United States, 1 in 5 people suffer from chronic bowel diseases such as irritable bowel syndrome. Seed is trying to develop a system that will help them by building a database of stool images. Seed’s research team will provide a tool that can diagnose user health status from images of stool so that users can manage their health.
Sid co-founder and CEO Ara Katz says stool isn’t usually considered a source, but stool is direct evidence of intestinal health. To cooperate with the establishment of the seed credit image database, enter the page to send credit images to the seed using a smartphone and press the button (#GIVEaSHIT) located at the bottom of the page. When the form opens, enter your email address and your usual credit hours. Then check the box and tap the upload photo to upload. The photographs taken are separated into personally identifiable email addresses and potential metadata and used by the Seed research team in an anonymous state.
Sid’s research team diagnoses and classifies all stool images through 7 specialists who specialize in digestive tracts who analyze stool images sent from the public. It is said that the classification consists of 7 types. These doctor-tagged stool images are used to train the AI model. The goal is to use an educational system, such as an autonomous vehicle learning to identify a tree or cat in the street, so that the artificial intelligence can check the stool condition like a doctor.
The purpose of the credit database creation project is to open source large credit images that are difficult to obtain in general to be useful for academic purposes. One expert working on the database construction project said that people with chronic diseases are contemplating choosing what to eat and how much exercise to choose. Say. Related information can be found here .