TEAM(119,VISUALVOX)

Inspiration

With the idea of "Inclusion" be created, many people are devoting to creating methods for those visual impaired to be able to surf the Internet. However, the biggest barrier for blind is the Images. They can not enjoy the image on any website. Some website add short descriptions, but tons of details were left out. How can we solve this to enable those blind to see "in detail", like a normal person with normal sight does?

What it does

This tool can cut a picture into serval sections by identifying "objects", and then it will make a dictionary, the keys are the name of the objects, and the values are the specific descriptions for the object. if the user wants more detail , just select that object, and the object would be cut into different sections to offer information in a deeper level. By a recursive process ,we successfully simulitate the process for a normal human to admire a picture.

How we built it

We use chatgpt4-vision to achieve image recognition. And I write the prompt and use FEW-SHOT and Chain of Thought to tell it the rules and boost its productivity.the programming language we use is python