I am working with my team on project given by one of the labs. Our task is to identify/detect objects in an image (centrioles and nuclei in our case). With the short period of time, the deliverable to complete for the lab and the report to write, we where wondering about the limitations of use of external libraries.
Is it allowed to use already created and trained neural networks if they are a good solution to our issue? If we cite correctly tthe material we use and precise how we use, it shoud not be illegal given the instructions of project 2. Is it also allowed to use a pre-existing NN and tweak it to solve our problem? Or do we have to code from the ground up this kind of solutions even though helpful materials that already exist elsewhere?
it is surely allowed to use any library or existing software you want, and also pretrained models. in fact a good 'related work' section in your report would point out what already exists. in any case, make sure to cite it properly, document things reproducibly, and make sure to also document your additional contribution on top of the existing approach
Authorized material fo project 2
Hello,
I am working with my team on project given by one of the labs. Our task is to identify/detect objects in an image (centrioles and nuclei in our case). With the short period of time, the deliverable to complete for the lab and the report to write, we where wondering about the limitations of use of external libraries.
Is it allowed to use already created and trained neural networks if they are a good solution to our issue? If we cite correctly tthe material we use and precise how we use, it shoud not be illegal given the instructions of project 2. Is it also allowed to use a pre-existing NN and tweak it to solve our problem? Or do we have to code from the ground up this kind of solutions even though helpful materials that already exist elsewhere?
Thank you for your response!
it is surely allowed to use any library or existing software you want, and also pretrained models. in fact a good 'related work' section in your report would point out what already exists. in any case, make sure to cite it properly, document things reproducibly, and make sure to also document your additional contribution on top of the existing approach
1
Add comment