Relic Finder

Uncover the Truth with Data.

Comment

Inspiration

Our interest in learning more about Computer Vision and AWS Rekognition motivated us to create a solution that identifies getaway vehicles. We aimed to explore how these advanced technologies could be applied to real-world problems

What it does

Our Project is designed to identify getaway vehicles involved in criminal activities. Utilizing AWS Rekognition, the model analyzes images to detect specific features and patterns that help pinpoint vehicles of interest.

How we built it

To bring our vision to life, we utilized AWS Rekognition to train and deploy our image recognition model, which processes vehicle images to identify potential matches. Our application leverages an S3 bucket to store and manage the images and results securely. The backend is powered by Python, using libraries like PIL for image processing, boto3 for AWS interactions, uuid for generating unique identifiers for each submission, and json for data management. We also created a website using React.js to display our results.

The development process included the following steps:

Model Training: The images were preprocessed and used to train the AWS Rekognition model to improve its accuracy in detecting suspicious vehicles.

Web App Development: We built a user-friendly web interface using React.js, Javascript and Tailwind.css

Testing: The application underwent rigorous testing to ensure reliability and accuracy, with feedback incorporated from potential users to enhance functionality.

Challenges we ran into

Data Quality: Ensuring high-quality images for training and testing the model was a challenge, as some images were unclear or poorly lit.

Integration Issues: Seamlessly integrating AWS services with our Python code required troubleshooting various API-related issues, especially with permissions and access.

Model Accuracy: Achieving a satisfactory level of accuracy in detecting suspicious vehicles took multiple iterations and adjustments to the training parameters.

Accomplishments that we're proud of

Successfully implemented a functional prototype that accurately identifies vehicles linked to the theft case.

Leveraged AWS services effectively to create a scalable and reliable application.

What we learned

The importance of data preprocessing and cleaning for improving model performance. Gained hands-on experience with cloud services like AWS and their APIs. Learned how to work collaboratively as a team to overcome technical challenges and deliver a cohesive project.

What's next for Relic Finder

Enhancements: Integrate more advanced machine learning models to improve detection accuracy and speed.

Expansion: Consider expanding the system to include other types of suspicious activities and incidents. User Feedback: Gather feedback from potential users to refine features and improve the overall user experience.

Scalability: Explore additional cloud services and features to enhance scalability and performance, such as incorporating real-time alerts or notifications.

Built With

Share this project:

Updates