3

I'm studying CS and this year I should obtain my bachelor's degree.

I'm required to do a mandatory internship, and I'm doing it with a professor that asked me if I would join his research project. Right now, my assignment is to search for a paper about a challenging task in computer vision and to implement it (prostate segmentation in TRUS images, without using supervised learning).

The problem is: I've only taken the first basic course in computer vision, and I don't understand at least 80% of every paper. What should I do? Should I tell my professor I'm not suited for this task? If I change the internship, and consequently, my thesis, I almost surely won't be able to get my degree this year.

2
  • 1
    It's not clear what you were told to do, but if you are having difficulty you should speak to the professor as soon as possible. Probably the internship is supposed to be hard, and you are supposed to keep working on it. Commented Apr 17, 2021 at 6:48
  • Don't forget to go to your university library and to read books (on paper). You can find good books on computer vision. I did borrow some at Paris Jussieu university library, and you could use OpenCV Commented Apr 21, 2021 at 5:10

3 Answers 3

6

You should definitely talk with your adviser about your concerns. Perhaps asking them to point you to the relevant material you should read before you start reading the papers, which in many cases are written for the experts in the field. Your adviser can estimate the level of knowledge you have according to your resume, so I doubt they will be surprised if you ask to help with finding out the right literature for you.

3

Clear and honest communication with your advisor is the key to complete your thesis. If you don't understand the papers you're reading, just honestly say this to him. He might be able to give you some hints or direct you towards the right direction. On the other hand, if graduating this year is not your first priority and you're really interested in other topics, you can switch to another internship.

During my Master's thesis in CS I learned the hard way that you won't look dumb to your supervisor if you just admit that you don't understand the papers, so don't be scared of that! Indeed, asking questions shows that you're actually doing something.

1
  • 2
    I agree with this answer, but just to be clear, it's important to say "I don't understand why they say X / how they did Y / how they obtained formula Z", rather than a generic "I couldn't understand this paper at all", otherwise the supervisor won't really be able to help.
    – Andrea
    Commented Apr 17, 2021 at 13:51
3

I had this problem when I first started in computer vision too - the solution ended up being to spend a month reading papers and textbooks until I understood how to implement the thing in question. It's a learning opportunity, just one that unfortunately comes with a high effort requirement. For what it's worth, if you don't understand the paper in question, work backwards via the references until you find something you do understand, and then try to figure it all out from the ground up. It's ok to tell the professor what you're doing to set expectations appropriately, incidentally.