remembrain

Inspiration

Dementia and Alzheimer’s disease affect millions of people worldwide. One of the most difficult challenges for patients is recognizing the people around them and remembering past interactions. This often causes anxiety and confusion during social interactions.

The inspiration for this project came from the idea that technology could assist memory in a gentle and non intrusive way. If a system could recognize the person standing in front of the patient and remind them who they are and when they last met, it could reduce stress and help patients stay connected with their loved ones.

This project aims to build an AI powered memory assistant that helps dementia patients recognize people and recall important relationships.

What it does

Remembrain is an AI powered memory assistant that uses computer vision to recognize faces and provide helpful reminders to dementia patients.

When someone stands in front of the camera, the system:

1. Detects the face using computer vision
2. Identifies the person using face recognition
3. Retrieves stored information about that person
4. Provides a reminder through text or voice

For example, the system can say:

"This is Rohit, your nephew. You last met him yesterday at lunch."

The goal is to reduce confusion and make everyday social interactions easier for people living with memory disorders.

How we built it

The first prototype was developed using a laptop webcam to detect and recognize faces in real time.

The system captures frames from the camera and uses computer vision models to detect faces. Each face is converted into a numerical embedding that represents unique facial features. These embeddings are compared with stored embeddings in a database to identify the person.

The project was built using:

Python for the main application logic OpenCV for face detection Face Recognition (dlib) for identifying people pyttsx3 for voice output JSON or SQLite for storing information about known individuals

Face recognition works by converting an image into a vector representation:

( f(x) = (e_1, e_2, e_3, ..., e_n) )

Where (x) represents the input image and (e_1 ... e_n) represent extracted facial features.

The similarity between two faces is calculated using Euclidean distance: If the distance is below a certain threshold, the system identifies the two faces as belonging to the same person.

Challenges we ran into

One major challenge was ensuring real time performance. Face recognition algorithms can be computationally intensive, especially when processing live video streams.

Another challenge was maintaining recognition accuracy in different lighting conditions and camera angles. To improve performance, multiple images of each person were used during the encoding stage.

Privacy was also a critical concern. Since facial data is sensitive, the system was designed to store recognition data locally rather than relying on external cloud services.

Accomplishments that we're proud of

We successfully built a working prototype that can detect faces through a webcam, recognize known individuals, and provide contextual reminders through voice output.

Another achievement was designing the system with future wearable integration in mind. The architecture allows the project to eventually run on smart glasses so that users can receive real time memory prompts without needing to interact with a computer.

What we learned

This project helped us understand how computer vision and artificial intelligence can be applied to assistive healthcare technologies.

We learned how face recognition models work, how facial embeddings are generated, and how to build a real time computer vision pipeline. The project also highlighted the importance of designing technology with empathy and accessibility in mind.

Beyond technical skills, we gained insight into how AI can solve real world problems related to aging and memory disorders.

What's next for remembrain

The next step for Remembrain is to integrate the system with wearable devices such as smart glasses so that dementia patients can receive reminders seamlessly during everyday interactions.

Future improvements include automatic tracking of the last interaction with each person, emotion detection to provide additional context, and secure encrypted storage for sensitive facial data.

The long term goal is to develop a fully functional AI memory assistant that helps dementia patients maintain independence and confidence in their daily lives.

Built With

• and-last-interaction-**laptop-webcam**-?-used-as-the-input-device-for-capturing-real-time-video-**computer-vision-models**-?-used-for-facial-feature-extraction-and-identity-matching-if-you-want
• computer
• computer-vision
• edge-ai
• face-recognition
• i-can-also-give-you-a-**much-stronger-?built-with?-section-that-hackathons-prefer**-(it-includes-buzzwords-like-**ai
• json
• numpy
• opencv
• python
• pyttsx3
• relationship
• sqlite
• vision