SMSshield

• 

  logo

• 

• 

• 

Why?

In an increasingly connected world, vulnerable groups like children and the elderly often become prime targets for scammers. These groups may not recognize fraudulent messages, leading to financial loss and emotional distress. Our mission is to bridge this gap and protect these individuals by offering a solution that not only warns them about potential threats but also educates them on how to handle such situations. By "connecting the unconnected," we empower them to use their devices with confidence and security. You can now rest easy, knowing that grandma or little Timmy will be able to identify and ignore the next scam SMS, such as "DHL Alert: Your package is pending delivery. Immediate action required! Visit dhl-urgent.com now to verify your details and avoid package return."

What?

Introducing SMSshield, a groundbreaking application designed to detect and flag spam and scam messages before they can cause harm. SMSshield scans incoming SMS messages for malicious content. When a suspicious message is detected, the app not only alerts the user but also provides educational tips and guidance on how to respond. This dual approach ensures users are both protected and informed, significantly reducing the risk of falling victim to scams.

How?

To build SMSshield, we conducted extensive research to gather data on common scam messages targeting vulnerable groups. Our development process involved:

• Technology Stack: Utilizing Kotlin, XML, and Java within Android Studio to create a robust and user-friendly application.
• Machine Learning Integration: Employing sophisticated ML models trained on a comprehensive dataset of scam messages to enhance detection accuracy.

Challenges We Faced

• Idea Development: Settling on the right idea took significant time, leading to a tighter development schedule.
• Design and Features: While core functionalities are implemented, some features had to be scaled back due to time constraints.
• Learning Curve: For some team members, this hackathon marked their first foray into new coding areas, which was both a learning opportunity and a challenge.
• Data Collection and UI: Collecting relevant data and designing an intuitive interface presented considerable hurdles that we overcame through intensive work in the final days.
• ML Models + Kotlin: Training ML models with python and then making them work in our Kotlin-App proved to be very time consuming, as for some models (lgb, xgb, cat) there was no onnx support.

Accomplishments

We are proud of:

• User-Friendly Design: Creating an interface that is accessible and easy to use for children or the elderly.
• Scam Detection Accuracy: Achieving high accuracy in certain areas of identifying scam messages, which has the potential to make a significant positive impact on users' lives.
• Functionality: Delivering a fully functional app ready for deployment, capable of effectively detecting scam messages.

What We Learned

Through this project, we gained valuable insights into:

• Technological Capabilities: Exploring the potential of Kotlin and XML for mobile app development.
• Teamwork and Collaboration: Building strong team dynamics and making new connections.
• Local Insights: Gaining a deeper appreciation for the city of Regensburg and its community.

What's Next for SMSshield

Multi-Language Support

To broaden our user base and ensure accessibility, SMSshield will be available in multiple languages, making it easier for non-English speakers to use and benefit from the app.

Advanced Machine Learning Models

We plan to replace the current Levenshtein distance method with more sophisticated machine learning models to improve detection accuracy and efficiency.

Community Reporting

Enabling users to report suspicious messages will help us expand our database and continuously improve our detection algorithms.

Emergency Contact Alerts

Implementing a feature that forwards dangerous messages to emergency contacts, ensuring that both children and seniors are protected through timely alerts to trusted individuals.

Full-Scale Messaging App

Developing SMSshield into a comprehensive messaging platform will allow users to manage all their communications within a safe environment, eliminating the need to use other potentially vulnerable apps.

By continually innovating and responding to user feedback, we aim to make SMSshield an essential tool in the fight against scams, particularly those targeting the most vulnerable in our society.

Built With

• android
• android-studio
• coding
• kotlin
• ml
• onnx
• python
• xml