Safety Pro

Make the internet your safe space

Comment

About The Project: Safety Pro!

Example image

Inspiration

Almost 75% of the users on the internet are between 12-20 years old. With the rise of cyber threats and inappropriate content on the internet, it's become imperative to ensure digital safety, especially for younger users. Recognizing this, we were motivated to develop a tool that could provide real-time protection against harmful content. Each day, countless individuals are exposed to potentially dangerous or unsuitable material without adequate filters in place. Our goal was to create a solution that not only blocks such content but also educates users about online safety.

What it does

Safety Pro is an innovative application designed to enhance internet safety. Once downloaded and installed via an executable file on our website, it continuously scans the laptop's screen using advanced AI and ML algorithms. The software identifies and blocks harmful content in real time, preventing it from being displayed. Users immediately get any and all harmful content blocked on their screen. This proactive approach ensures a secure and educational browsing experience.

How we built it

We wanted to ensure that our machine learning model was extremely accurate keeping in mind the crucial information that it was predicting. We wanted to make sure that our website was easily navigable and users would have no trouble going through our website. We chose colors that were inviting and would ensure that users have a positive experience going through our pages. Our front end was created using Javascript and our Backend was created using Python. We also used different api's for our backend machine learning models. Features:

  • Webpage Scanning: Our code takes a real time feed of the user's screen and for each time the user is working or viewing something it makes sure it is observing in real time what is going on with the user's display.
  • Image To Text Mapping: Our code also simultaneously views the user's feed images and highly accurately converts the text available on the screen into text that can be manipulated and tested for its sentiment.
  • Text Sentiment Analysis: Whatever Text is available to view on the users screen is then run through a highly sophisticated machine learning model that predicts with a very high certainty whether the text can be deemed harmful or not.
  • MultiThreading: Our code also utilizes multithreading in order to ensure that we can work highly efficiently due to the nature of our task, ensuring that we can display our work in real time.
  • Blocking Inappropriate Content: Based on our text classification using Machine Learning and Natural Language Processing, any and all text that we do find that is considered harmful is blocked from the users screen until it remains on the screen, ensuring that the user is protected from harmful content, making their time on the internet be a safe space for them.

Challenges we ran into

  • It was definitely really difficult to try to get our project to work close to real time working with the vast amount of classification we had to do.
  • Incorporating multi threading in 24 hours was a tough task to accomplish, it took us a while to ensure our logic was correct for our code, ensuring that we were able to seamlessly hide all NSFW content.
  • Finding data to train the model on was also challenging as we were trying to tackle an issue that does not have a lot of research or prior work done on it in the machine learning field.
  • It was also tough for us to fine tune an open source model to make sure it was working well for our intended use case for this specific scenario.

Accomplishments that we're proud of

We are particularly proud of Safety Pro's ability to deliver real-time, accurate content blocking without hindering the user experience. Achieving a balance between performance and functionality, especially in a tool designed for continuous background operation, was a significant milestone. It was also really nice to see our project working, our project involves a lot of moving parts behind the scenes and being able to do multiple hard parts and see this project work was definitely an accomplishment the entire team was proud of.

What we learned

As a diverse group, we split the tasks amongst ourselves according to each member's skill. By doing so, we learned how to communicate effectively with each other and share data more efficiently. We also learned frameworks that we had not used before like JavaScript and Python. We had to learn how to use API's and how to effectively communicate using our front end and back end frameworks.

What's next for Safety Pro

Moving forward, we aim to expand Safety Pro’s capabilities to include more nuanced content detection, such as detecting harmful audio in videos as well as detective images that are also NSFW. We also plan to develop a mobile version to provide comprehensive protection across all devices. We also want to improve our application even further to ensure that it is even faster than it currently is, blocking all NSFW content as fast as possible.

Share this project:

Updates