Enrollment is the process of detecting faces in an image (or video frame) and creating “templates” that encode the identifying characteristics of each face. The faster the enrollment speed, the less computing power required to support a face recognition application.
The template size is the amount of storage space required to save a template extracted in the enrollment process. Applications involving the search of a database of face records generally must cache templates in memory; thus the smaller the template size, the less RAM needed.
The comparison speed is how long it takes to measure the similarity between two templates. The faster the comparison speed, the faster the verification or search results.
While AI has become a popular term in business circles and pop culture, it generally lacks any specificity as to what techniques are being applied to a given problem. In many ways all face recognition algorithms developed in the past two decades were AI. However, the recent rise of convolutional neural networks and deep learning frameworks have unlocked the means to learn representations that are vastly superior to the manual representations of the past.
In terms of Rank One's intellectual roots, they are more properly described by the academic disciplines of pattern recognition (see Duda and Hart, 1973) and computer vision (see OpenCV). Rank One's team of engineers have matriculated under and collaborated with some of the leading minds in these fields, and our software bears the marks of the parsimony and precision such disciplines extol.