In this project, we try to make a model that detect maize and weeds in a photo taken from 1 meter, vertically to the ground. In fact its enough that model predict it is MAIZE or NOT. But Weeds can be "quite different as morphologically".
My opinion is that we must annotate each species as a different classes. But I still need someone say that its a obligatory.
(Most probably, we use yoloV5 or V8)
Thank you.
You can annotate the dataset with 1, 2, 3 or any number of classes. Yolo is capable of performing accurate detection on single class and binary classes. For single class you will get AP equal to MAP. Whereas, in more than one class problem, you will get AP for each class and MAP for the entire test set. AP helps in identifying which particular class is detected with high precision.
I hope this helps you
Further, it depends on you whether you want to focus on Maize only or consider Weed as well. In case you want to detect Maize only then perform annotations for it only. If you want to detect Weed as well, then you are suggested to perform annotations for both the classes. It will reduce false positives since Weed can also be detected as Maize as the bounding boxes are going to be the same for both the classes
How would anyone else answer this question?
As Akhil Kumar mentioned, Yolo is obviously capable of doing multi-class classification/segmentation. But what you're asking is part of your study design. This is something you and your team or advisor have to decide imo.
- Badges
- Science topic
- Similar topics
- Correction