An empirical study of dermatoglyphics fingerprint pattern classification for human behavior analysis

Human measures many things consciously or subconsciously by touching and sensing without any measurable challenges. However, measuring intangible features like human behavior is a challenging task. Human behavior analysis is an important computer vision technique with a lot of attention which includes human–computer interaction and assisted living. It is essential not only for a wide range of applications but also to understand the abnormal behavior of humans. This survey paper reviews various methods used for the analysis of human behavior by dermatoglyphics fingerprint pattern classifications. This survey studies 50 research papers based on fingerprint pattern classification and presents techniques related overviews such as deep learning (DL)-based methods, convolutional neural network (CNN)-based DL methods, machine learning (ML)-based methods, Support vector machine (SVM)-based ML methods, and optimization methods. The overview of this survey comprised of classification of research methods, year of publication, evaluation metrics, employed datasets, and toolsets for human behavior analysis. The analysis demonstrates that accuracy is the most commonly used evaluation parameter in fingerprint classification which is used by 33 research papers. Finally, the research gap of analyzed methods is explained, which encourages researchers to develop new effective methods for human behavior analysis using fingerprint pattern classification.