MindBloom
Your memory assistant
The problem MindBloom solves
🧩 Problem It Solves
Caring for individuals with Alzheimer’s and dementia is extremely challenging. Patients often face memory loss, confusion, wandering, and difficulty managing daily routines. Caregivers struggle with constant monitoring, emotional stress, and the need to provide consistent reminders and safety oversight.
There is no unified, intelligent system that supports both the patient and the caregiver in real time — until now.
NeuroLink addresses this by offering an AI-powered cognitive assistant, a caregiver dashboard, a patient-friendly mobile app, and immersive VR experiences. It keeps patients oriented, safe, and emotionally supported, while giving caregivers real-time insight, control, and peace of mind.
🚀 What People Can Use It For
For Patients (Mobile App):
-
AI Memory Assistance
Automatically recognizes familiar faces stored in the journal and reminds the patient who they are in a gentle, supportive way. -
Routine & Medicine Alerts
Sends alarms when medication is missed, helping ensure timely dosage and preventing health risks. -
Daily Guidance & Emotional Support
Offers simple, voice-based interaction and grounding prompts to reduce confusion and stress. -
VR Memory Recall Therapy
Allows patients to view AI-generated memory videos in an immersive VR environment for cognitive stimulation and emotional comfort. -
Safety & Orientation
Provides location tracking and contextual guidance to help patients navigate daily life with greater independence.
For Caregivers (Web Dashboard):
-
Real-time Monitoring
Track medication adherence, patient location, mood indicators, and app activity. -
Live Alerts
Receive instant notifications if the patient misses a medicine dose, wanders, or requires help. -
Memory Journal Management
Upload photos, videos, and notes to train the AI to recognize important people and moments. -
Routine Management
Set reminders, schedules, and personalized safety parameters. -
VR Content Creation
Generate personalized AI-powered memory journey videos that the patient can later experience in VR.
How It Makes Life Easier and Safer
- Reduces caregiver burden through automation of reminders, monitoring, and alerts.
- Improves patient safety with real-time notifications and wandering detection.
- Enhances emotional well-being by helping patients recognize loved ones and revisit meaningful memories.
- Supports cognitive function through AI-driven reminders and VR therapeutic sessions.
- Accessible anytime via a patient mobile app and caregiver web dashboard, making it easy to use for families and professionals.
Challenges we ran into
Challenges I Ran Into
1. Face Recognition Accuracy for Alzheimer’s Patients
One major challenge was ensuring the mobile app could correctly recognize people stored in the journal. Alzheimer’s patients may have older photos or inconsistent lighting conditions, which caused the model to misidentify faces.
To solve this, I added preprocessing steps such as image normalization, face alignment, and fallback prompts when confidence scores were low. This greatly improved recognition accuracy and reliability.
2. Synchronizing Mobile App and Web Dashboard Data
Real-time updates between the patient’s app and the caregiver’s dashboard were initially inconsistent due to delays in database writes and API responses.
I addressed this by implementing real-time listeners and WebSocket-based updates so that alerts, journal changes, and routine edits synced instantly across devices.
3. Integrating VR With AI-Generated Memory Videos
Displaying AI-generated videos inside a VR environment was tricky, especially since different browsers handle WebXR differently. Some devices couldn’t load the video textures smoothly.
The fix involved optimizing video resolution, preloading assets, and using a more compatible WebXR player library that ensured smooth playback across multiple VR-capable browsers.
4. Missing Medication Detection Logic
Creating an alarm that triggers only when a patient actually misses medication (and not just closes the app) was challenging. The initial logic raised false alerts.
I solved this by building a timestamp-based confirmation system: the patient must actively confirm taking the medicine, and the system cross-checks this with reminder schedules before sending an alert.
5. Handling Security for Sensitive Health Data
Because the system stores personal memories, identity data, and location logs, ensuring privacy was a big concern.
I implemented role-based access, encrypted storage, and secure API tokens to make sure only authorized caregivers could view or modify sensitive information.
Tracks Applied (1)
Web2.0 Track
Technologies used