Skip to content

li-pearl/nemo

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

75 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

NEMO

Wearable AI Facial Recognition Assistive Device For Dementia and Alzheimer's Patients

Regeneron ISEF 2023 - ENBM084

Rationale

As of this year, there are over 6 million Americans living with Alzheimer’s disease, with the number projected to more than double by 2050. The number of people with various dementias worldwide is also expected to rise from the current 55 million to 139 million by 2050. Alzheimer’s and dementia has been proven to impair the facial recognition abilities of patients, a form of declarative memory. This leads to trouble recalling or forgetting entirely the faces and names of loved ones, depending on the severity of the disease. Alzheimer’s affects the ability to connect faces to names because of the complexity and similarity between different faces and the arbitrary association to names. This decline can seriously affect the social life and confidence of people living with dementia as they feel they are no longer in control and can’t trust their own judgment. However, it has been found that in early-mid stages of Alzheimer’s and dementia, patients often don’t entirely forget faces and names, but have difficulty recalling the information on command leading to frustration and social withdrawal. Many care facilities have introduced the use of picture boards or other visual aids to increase face-name memory and quality conversations. Although the efficacy of such reminding methods declines as the disease reaches the later stages, an affordable, compact, and discreet device that serves simple reminders of loved ones’ names could drastically increase the confidence of people with early to mid stages of dementia by allowing them to take back some control. Such a device could use Artificial Intelligence (AI) facial recognition technology to give verbal reminders of the names of loved ones through a discreet earpiece. A database of loved ones’ and caregivers’ pictures and names will be uploaded to the device which will use a tiny off-the-shelf camera that, when activated by the user, will use AI facial recognition to query the database, match the face to a name, and enunciate it. To keep accessible, the device will be open-source, affordable, use off-the-shelf components, and use 3D printing for custom parts. Even though there is currently not much on the market when it comes to using facial recognition as assistive technology for people with dementia, there are currently a couple of applications made to be run on mobile devices that can recognize faces and display their respective names. While useful, this requires opening an app, pointing a phone or tablet camera, and reading the text on the screen, something the user may be uncomfortable to do in public. This project aims to be “out of the way” in order to not break the natural flow of conversation, and therefore, increasing social confidence amongst users.

Required Packages

  • face_recognition
    • wrapper for dlib
  • opencv
  • cmake
  • pyttsx3 version 2.71

Windows: Using python 3.7.6

Raspberry Pi: 3.7.3

Hardware

  • Raspberry Pi 4B
  • PiCamera V2 module

Installation

Competitions

  • 2022-2023 Pennsylvania Junior Academy of Science (PJAS) Region 4
    • Qualified but not competing at States
  • 2022-2023 Capital Area Science and Engineering Fair (CASEF)
  • 2023 PA BioGENEius Challenge
    • Healthcare Challenge (Medical Biotechnology)
  • 2023 International BioGENEius Challenge
    • Healthcare
  • 2023 International Science and Engineering Fair (ISEF)
    • Biomedical Engineering (ENBM084)
  • Hackaday Assistive Technology Challenge

Awards

  • 🥇1st Award - PJAS Region 4
  • 🥇Grand Champion - CASEF
  • 🥇PA Biotechnology Insititue BioGENEius Challenge Winner
    • Healthcare Challenge
  • Competing at 2023 International BioGENEius Challenge - May 19-22
  • Competing at PJAS States May 14-16 ISEF Conflict
  • NC State College of Engineering Award - 2023 Regeneron ISEF

Acknowledgements

  • 💸 Susquehanna IEEE, Lemelson Foundation, CASEF and PJAS for providing funding and encouragement to continue making
  • My grandfather
  • Open source example contributers in face_recognition

Dedicated to my boro ma

1919-2023

Releases

No releases published

Packages

No packages published

Languages