The Bachelor thesis entitled “MPAIFR: Missing persons age invariant face recognition” addresses a significant and constantly evolving challenge in the field of biometrics: that of identifying identity, despite physical changes induced by time. Security systems, card verification, video surveillance, credit criminal identification, person identification [1], passport renewal, law enforcement, and bio-metric authentication [2] are just a few of the numerous applications that employ face recognition technology. There is often the need to recognize and identify missing persons by comparing newfound images that could have been taken years apart. There may be cases when the forensics experts are required to change the age of the person's face. In the current times, the ability to identify missing persons, no matter how many years they have been missing, represents a significant advancement. Hence, this demands a robust age-invariant face recognition model in the presence of aging.
One of the primary obstacles facing AIFR is the large disparity brought about by aging, but also by gender. We human beings can discriminate between males and females accurately based on face, the pitch of voice, the difference in gait, behavior, etc. [3]. Recognizing faces across aging is difficult even for humans; hence, it presents a unique challenge for computer vision systems [2]. The paper is centered on deep learning models, underpinning age-invariant facial recognition systems. Several topics are presented, ranging from image preprocessing, to compensate for variations in illumination and position, to techniques that model and adapt facial features according to age and gender. The ability to learn and represent hidden features from the given input data has made deep learning models successful in achieving better results in contrast to the traditional conventional approach. Specific challenges are also presented, such as the lack of appropriate training data for missing persons, particularly, over extended periods, and how emerging technologies, such as transfer learning and data augmentation, can help to surmount these.
By considering all of these aspects, the work seeks to contribute, to the development and implementation of age-invariant face recognition. This not only advances the field but also contributes to reuniting those who are lost to their families, opening up new horizons for the applicability of biometrics, within society.
[1] M. Yasmin M. A. Shahid A. Rehman M. Sharif, F. Naz. Face recognition: A survey. Journal of Engineering Science and Technology Review, 10:166–177, 06 2017.
[2] K. M. Bhurchandi M. M. Sawant. Age invariant face recognition: a survey on facial aging databases, techniques and effect of aging. Artificial Intelligence Review, 52:1–28, 08 2019.
[3] Indiramma M. Nayak, J. S. An approach to enhance age invariant face recognition performance based on gender classification. Journal of King Saud University - Computer and Information Sciences, 34:5183–5191, 2022.