2019/05/17 by Azin Asgarian, Asgarian, Azin, Shun Zhao +14 · 1 citation
Computer Science · Medicine · Neuroscience · Psychology · #Artificial Intelligence (cs.AI) #Body Image and Dysmorphia Studies #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Face Recognition and Perception #Face recognition and analysis #Facial Nerve Paralysis Treatment and Research #Machine Learning (cs.LG) #Machine Learning (stat.ML)
paper · pdf · doi:10.48550/arxiv.1905.07446
openalex publication_date 2019/05/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Accurate facial expression analysis is an essential step in various clinical\napplications that involve physical and mental health assessments of older\nadults (e.g. diagnosis of pain or depression). Although remarkable progress has\nbeen achieved toward developing robust facial landmark detection methods,\nstate-of-the-art methods still face many challenges when encountering\nuncontrolled environments, different ranges of facial expressions, and\ndifferent demographics of the population. A recent study has revealed that the\nhealth status of individuals can also affect the performance of facial landmark\ndetection methods on front views of faces. In this work, we investigate this\nmatter in a much greater context using seven facial landmark detection methods.\nWe perform our evaluation not only on frontal faces but also on profile faces\nand in various regions of the face. Our results shed light on limitations of\nthe existing methods and challenges of applying these methods in clinical\nsettings by indicating: 1) a significant difference between the performance of\nstate-of-the-art when tested on the profile or frontal faces of individuals\nwith vs. without dementia; 2) insights on the existing bias for all regions of\nthe face; and 3) the presence of this bias despite re-training/fine-tuning with\nvarious configurations of six datasets.\n