Catfishing Prevention

Inspiration

Catfishing has affected 27% of online daters in the past 12 months. In 2022 in the US there were 19 050 romance scam complaints causing $739 million dollars in losses. In these cases, many people would fall into the trap and start a fake date with the scammer. In the end, most of them would be scammed with their money and sometimes private information. Many people dating online have encountered many cases of catfishing and many have gotten scammed. The scammers would often use AI generated pictures or online pictures to disguise their real identities. In the end, most of them would be scammed for their money and sometimes private information and photos.

What it does

This app can detect if the image of the person is from the web. In the future it will also compare names of dating profile to internet match or check for AI generated images if there is no match found.

How we built it/How it works

We built this prototype using DeepFace and its accompanying libraries/frameworks and Streamlit for the User Interface. This works by receiving a image from users that then allows DeepFace to compare the input image to every face in its database (local), giving each one a similarity score from 0-1. 0 would be a perfect match and 1 would be very dissimilar. A score less than 0.4 was considered as the same person and if two scores are present, the lower score will "win" and be detected as the match.

Challenges we ran into

The use of DeepFace and Streamlit was not too challenging as they are prebuilt python libraries. The main challenge was learning how to use them through many internet searches and downloading all the necessary libraries to make python work on Visual Studio and DeepFace work alongside. We had a lot of troubles with some of the newest versions not being supported and needing to reinstall to connect software.

Accomplishments that we're proud of

This was our first hackathon experience and we are very proud of how we were able to complete a project that forms the base of a more in-depth catfish identification software. We were able to identify the necessary Python libraries and apply them to create a usable UI and a working back end.

What we learned

We learned about different Python libraries that we can use to code this type of application. We deepened our understanding of Python, going from very surface level to being able to dig a little deeper. We learned how to use DeepFace and especially Streamlit to create the basic prototype that we have.

What's next for Catfishing Prevention

The next steps for our Catfishing Prevention is to connect our program to an Image API such as Tineye API to widen our database to almost the whole internet. From here, we want to add an additional program that if a match is found, will compare the names of the people in each image to see if they match. If there is no match, we want to add AI image detection that will flag the other possibility of a cat fisher.