I don’t think you need to do that. for a scientific paper, citing OpenCV is plenty. you don’t have source code you are distributing. even for a slide set/presentation, a tidy citation under the slide should be enough. the computer vision research community “steals” slides from each other all the time (with credit).
don’t be surprised, OpenCV likes to get cited with the original release from around the year 2000, but you should also note what version you are referring to.
practically, there is a difference between science and commercial use. in science, it’s about abiding by academic standards, and citing fulfills that because you declare the cited work to be someone else’s, giving credit. licenses are mostly intended to regulate commercial use, i.e. determine if companies can make money from the library, and if they have to mention the library, …