Project Summary
A Novel Aid with Facial Detection, Facial Recognition, and Audio Analysis Capabilities for Patients with Dementia or Communication Disorders and their Caretakers

a usually progressive condition marked by deteriorated cognitive functioning often with emotional apathy.

Communication Disorders:
...a wide variety of problems in speech, language, and hearing

The number of people diagnosed with dementia-causing diseases is rising at an alarming rate. Currently, the medical devices that exist for such diseases mainly serve medical purposes and basic task handling. It is imperative to create medical devices that will improve the quality of life and independence for patients.

Communication disorders and often Alzheimer’s disease interfere with social interactions. Caregivers and guardians who cannot be near the people with special needs at all times may face the challenge of understanding events in their days without being present. A source for a direct account of their day would aid caregivers and guardians in better helping people with special needs.

I developed a system that collects the important events in a day—recognized faces and loud audio—and combines them into a concise record for patients or caregivers to review at the end of their day. 

The developed system uses Haar-like features to detect a face and Eigenfaces, based on principal component analysis, to recognize the face. Through a supervised process, the system is able to recalculate the Eigenfaces and become more accurate. At the same time, audio segments are captured and analyzed for audio pressure. The system is able to identify the important events and compile the data into a concise, chronological log, which displays the image captured showing the recognized person and his/her surroundings, a “known” image of the person, the person’s name, the time and date of the encounter, and the audio recording.

I tested both the accuracy of the facial recognition in my system, and the ability of the system to distinguish important events and compile the log. Not only did the results show the compilation capabilities to be 100% accurate, but also that the facial recognition is on average 85% correct after one learning process.

The results of the experiment supported the hypothesis that by enabling a video-audio recorder with decision-making abilities based upon facial detection, facial recognition, and audio analysis, an aid for patients and caretakers can be created that compiles a concise record of a day’s important events.

The plan is to make the system into an application for portable mp3 players and smartphones, which will make the system less expensive and more accessible. The devices will be worn around the neck of the patient.

Concise Log System


Project Summary
Fill out the project summary
 and be sure to include your video or presentation.


Judges' Tips An excellent film or presentation will provide a clear, brief overview of the question you are investigating, the stages of your project, what you set out to achieve and how far you succeeded.