The OIDv4 ToolKit is a neat Python package that lets you syphon off parts of Amazon's OpenImages dataset into custom object detection datasets of your own. At Roboflow, we often merge these with our own datasets (in conjunction with the label remapping pre-processing step) to get more training data.
Unfortunately, the format that OIDv4 spits out isn't compatible with any known models -- you'll need to convert them to a standard format with a tool like Roboflow.
Below, learn the structure of OIDv4 TXT.
helmet 0.0 73.041386 355.2 463.235188
person 2.56 95.46632500000001 86.4 213.357589
helmet 123.52 26.9102 826.24 594.5815520000001With Roboflow supervision, an open source Python package with utilities for completing computer vision tasks, you can merge and split detections in OIDv4 TXT. Read our dedicated guides to learn how to merge and split OIDv4 TXT detections.
Below, see model architectures that require data in the OIDv4 TXT format when training a new model.
On each page below, you can find links to our guides that show how to plot predictions from the model, and complete other common tasks like detecting small objects with the model.