Ingredient Label Scanner

The ingredient's description page for a cleaner product.
This shows the backend of how the Optical Character recognition software recognizes the text.

Inspiration

The majority of cleaning products in the United States contain harmful chemicals. Although many products pass EPA regulations, it is well known that many products still contain chemicals that can cause rashes, asthma, allergic reactions, and even cancer. It is important that the public has easy access to information of the chemicals that may be harmful to them as well as the environment.

What it does

Our app allows users to scan a product's ingredient label and retrieves information on which ingredients to avoid for the better of the environment as well as their own health.

How we built it

We used Xcode and Swift to design the iOS app. We then used Vision for iOS to detect text based off a still image. We used a Python scrapper to collect data from ewg.org, providing the product's ingredients as well as the side affects of certain harmful additives.

Challenges we ran into

We had very limited experience for the idea that we have in developing an iOS app but, we wanted to challenge ourselves. The challenges the front end had were incorporating the camera feature and the text detector into a single app. As well as trying to navigate the changes between the newer version of Swift 11 from older versions. For our backend members, they had difficulties incorporating databases from Microsoft/Google, but ended up using JSON.

Accomplishments that we're proud of

We are extremely proud of pushing ourselves to do something we haven't done before. Initially, we had some doubt in our project because of how difficult it was. But, as a team we were able to help each other along the way. We're very proud of creating a single app that can do both a camera feature and Optical Character Recognition because as we found out, it's very complicated and error prone. Additionally, for data scrapping, even though the HTML code was not consistent we managed to successfully scrap the necessary data by taking all corner cases into consideration with 100% success rate from more than three thousand HTML files and we are very proud of it.

What we learned

Our teammates working in the front end learned how to use Xcode and Swift in under 24 hours. Our backend team members learned how to scrap the data from a website for the first time as well. Together, we learned how to alter our original expectations of our final product based on time constraint.

What's next for Ingredient Label Scanner

Currently our project is specific to cleaning products, however in the future we would like to incorporate other products such as cosmetics, hair care, skin care, medicines, and food products. Additionally, we hope to alter the list of ingredients in a more visual way so that users can clearly understand which ingredients are more dangerous than others.