In the field of facial recognition, achieving temporal invariance represents a very significant challenge. Facial recognition technology has demonstrated remarkable abilities to identify people at various life milestones thanks to the recent developments in machine learning. Nevertheless, the task is to synchronize infants and children with their adolescent counterparts, as facial features undergo constant changes at this stage. In this project, newborns and toddlers relate to their future adult facial profiles using the latest open-source face recognition algorithms. Although the accuracy is very high in adult facial recognition, the morphological changes in the children pose some unique challenges. Remarkable strides in the performance and accuracy of face recognition technology have largely been directed towards adults, with a notable gap in the development of systems specifically designed for children and sub-adults. Most existing face recognition systems are tailored for adult faces, leaving a limited scope for the application of these advances to younger age groups. Children are always in a state of growth [1][2][3], and with respect to face recognition, this is their ever-changing youthful appearance. Physical and textural changes of the face in children baffle current technology [4]. The development of newborns and toddlers is characterized by quick facial changes such as the changes in proportions, the formation of expressions, and the evolution of features. It provides a great challenge in the training of highly accurate models based on the limited datasets for this specific demographic. This research aims to assess the viability of recognizing faces from childhood to adulthood by evaluating verification performance. Verification, one of the two face matching methods, is considered less challenging as it involves comparing only two face images to determine if they match or not. To address these challenges, this study utilizes open-source algorithms and extends the application of facial recognition to encompass the intricate developmental stages from infancy to maturity.

Recommended Citation: Goli A., Ricanek K., Modaresnezhad M., ElSaid A., (2024). Face Recognition: How Easy is it to Match Newborns or Toddlers to their Adultselves. UNCW MS CSIS Proceedings. V. 18 , N. 9 .