Automatic dataset generation and recognizer training for recognition of subtle facial expression changes for smiling faces

Abstract

Recognizing the intensity of facial expressions in daily life is crucial to estimating the individual's Quality of Life (QOL). To recognize ambiguous expressions, a two-input facial expression recognizer specialized for individuals has been proposed in our laboratory. Although this recognizer can accurately recognize facial expressions based on comparisons, it has the limitation of requiring the manual provision of supervised data for each individual. The goal of this study is to automatically generate datasets required for training the two-input facial expression recognizer for smiles from videos that capture the facial changes of an individual. We propose a method that utilizes a conventional facial expression recognizer to extract data on explicit expression differences. Additionally, we introduce a three-input facial expression change recognizer to extract facial images with subtle expression differences. Experimental results showed that our proposed framework could successfully generate a dataset for training the two-input facial expression recognizer.