The objective of this research is to automatically detect the intake of meals for elderly people with dementia living alone by using a Pose Estimation procedure with TensorFlow and OpenCV.
 Such service based on an Artificial Intelligence product will require minimum intervention of the caregiver or a person as medical support.
 
 We proposed a method for the automatic eating activity detection. We will use this approach for human activity detection in general, for instance monitoring the security and the protection of the patient, automatic motivation of the patient to eat in case no eating detection has been done.
 The choice of appropriate AI assistive technology was done to satisfy both the elderly people with neurodegenerative disorders and the caregiver, to verify the ethical aspect, simplify design, optimize code, and improve user friendly aspects.