FishBodySet4 Computer Vision Project

LingWei

Updated 6 months ago

58

views

5

downloads
Description

Hi!

There should be two categories: dot and fish.

We will label them by boxes (the second tool you may find on the right-hand side). We want the boxes to include all the pixels of the objects. So it's OK to get the boxes slightly larger (e.g. one more two more pixels wider) instead of having boxes too small so that we would miss some pixels. Please make sure that the entire fish head, fish tail, and fish fins are included in the boxes.

Sometimes, the dot (shadow) can overlap with a fish, or two fish can overlap with each other. It is OK to have overlapping boxes. Try to recognize both objects that are overlapping with each other, and draw a box for each one. If it is really very difficult to identify the fish inside a dot, it is OK not to label it.

Some tricks: (1) When you select the bounding box tool, press Ctrl without releasing it, and your cursor becomes the drag tool, and vice versa. (2) Try to zoom up a bit before you draw the boxes.

Supervision

Build Computer Vision Applications Faster with Supervision

Visualize and process your model results with our reusable computer vision tools.

Cite This Project

LICENSE
CC BY 4.0

If you use this dataset in a research paper, please cite it using the following BibTeX:

                        @misc{
                            fishbodyset4_dataset,
                            title = { FishBodySet4 Dataset },
                            type = { Open Source Dataset },
                            author = { LingWei },
                            howpublished = { \url{ https://universe.roboflow.com/lingwei/fishbodyset4 } },
                            url = { https://universe.roboflow.com/lingwei/fishbodyset4 },
                            journal = { Roboflow Universe },
                            publisher = { Roboflow },
                            year = { 2024 },
                            month = { mar },
                            note = { visited on 2024-09-27 },
                            }
                        
                    

Similar Projects

See More
1.4k images
111
1.4k images
22
1.4k images