The text below is written by me, and might not reflect the views of the department.
If you see me (or another faculty member) listed for a section of CSc 4999, CSc 4982, or similar classes, it probably does NOT mean what you think it does. If you see a faculty member listed, it most likely means that he/she agreed to offer this for a student meeting the above criteria last year. It does not mean that he/she would be willing to do this again this year. Also, if a faculty member is not listed, he/she might be willing to do this. Any faculty member could have a section of one of these courses, and adding a new section is easy.
I have agreed to conduct these courses in the past, and I might do so again in the future. Courses like this, from my point of view, are a drain of my resources without any recompensation. I consider them to be a favor; if I know who you are, I might be willing to supervise your work, but keep in mind that you are asking for a favor, and I might say no.
So, what do you need to do to take such a course? You should make sure that you meet the requirements first. Then you should form an idea of what you would do in such a course. If you took a 4000-level classes, did very well, and can think of an extension to that material, especially one that fits with the faculty member's research, then you should approach him/her. It is important that you know what the faculty member's research interests are, and if you do not, you are not prepared to have a conversation about these courses. Depending on the faculty member, there may be an on-going project that you could join.
However, there is another element here: independence. You should manage yourself, and you should use your time wisely. Your advisor does not have the time and energy to watch over your shoulder to make sure that you do the reading. If your advisor tells you to read a paper, and you immediately think "he is not going to know if I read this or not, so I won't do it", then this will end badly. With that said, you might start to read the paper, and realize that it does not really help. In that case, you should put it aside and focus on one of the other references, then talk about it with your advisor. This is how research goes: you get a bunch of references that all look promising, you start reading them, and you find that some are excellent, some are OK, and some are simply not useful. After a while, you get very good at deciding which category a reference falls into.
2. There is a project or experiment to perform
Speaking of independence, I use the term "advisor" above, because "supervisor" does not really fit. I look at research work like this differently than a "boss/worker" relationship. I don't want to tell you what to do and have you do it mindlessly. Instead it's more like a "captain/navigator" relationship, where you're the captain. To go with that analogy, the research project is the ship, and you make the final decisions. I point us in the right direction at the start, help you avoid obstacles, and suggest where to go next. If the ship runs aground (i.e., you quit putting in the time required), I might have to relieve you of your command. I'm not going to get on-board with your idea if I think it will end like that, however.
3. There are regular meetings, at least once a week, maybe more
Don't say that you don't come to campus on the day that works for me to meet. Remember, this is a favor. If someone agrees to lend you money, it would be very poor form on your part to insist that it be all in ones. Likewise for research meetings, you are not as busy as you think you are, and you are certainly not as busy as your advisor. Missing meetings, or being late, is a sign that you do not take it seriously. Yes, things happen to people, so missing one meeting due to an unexpected problem is understandable. Having to cancel a meeting in advance is also understandable. If you cannot make a meeting, you should let me know right away. If you let me know less than 24 hours that you cannot meet, this is a problem. If that happens more than once a semester, you have a problem, and I might withdraw you from the course. Find a way to make it work.
4. There are some deliverables
This could be a presentation (with slides), or a demonstration (where you talk about your project, then run the code or show the hardware working), or a video (like a presentation with demonstration recorded in advance), or a poster show, or a paper (report, conference paper, etc.). It may be several of these things. Your audience might be just your advisor, or a research group, or a class, or even open to the public.
You will have deadlines to meet, such as the end of the semester. You might need to develop your own time-line to make sure that you meet the deadlines. The deliverables might have format requirements. If you are working with me, anything written should be done in LaTeX, and conform to the IEEE style for citations and bibliography/references.