Method for Emotion Corpus Validation from the Consensual Identification of Patterns in Alzheimer’s Patients

The present research proposes a method for the construction of a corpus for the early detection of Alzheimer’s, by identifying basic patterns of emotions on video, from the collection of patient information in the form of videos, analysis and identification of emotions based on facial expressions and finally validation by two statistical measures: weighted Kappa and Kappa. Applying the method on a corpus of 40 videos, an average score of 0.60 obtained in the Kappa index and 0.67 in the weighted Kappa index, which indicates a good agreement among the observers, and provides encouraging results for the use of the corpus in automatic learning, on the patterns of emotions that allow the detection of Alzheimer’s disease.