| Abstrakt: |
In recent decades, there has been a significant focus among researchers and developers on the issue of facial recognition as well as designing corresponding algorithms. These algorithms have shown substantial benefits when applied to a variety of industries, including video surveillance, criminal identification systems, building access control, and unmanned autonomous vehicles. In pursuit of efficient face profiling techniques from an image analysis perspective to detect critical features through local, holistic or hybrid methodologies, amongst others, research continues with particular attention towards reviewing outstanding methods within each approach while also developing standardized categorization criteria. The investigation provides a comprehensive assessment of multiple strategies by comparing their respective strengths and limitations based on robustness, accuracy intricacy level, and discriminability measures. The research consists of an intriguing aspect related to the database utilized for facial recognition. The study covers both supervised and unsupervised learning databases that are commonly used, elaborating on numerical outcomes achieved by the most viable methods, along with a summary of experiments conducted and challenges faced in implementing these techniques. Moreover, notable attention is given to determining prospects for future face recognition research and development projects through thorough analysis. [ABSTRACT FROM AUTHOR] |
