All images in Supervisely have metadata by default. It contains base information about image in dataset such as: time when image was created or modified by user, pixel area, image size, resolution. But there's also an ability to store your own custom metadata for an image. In this example we store some additional information about image, which helps to match images from one database with another, such as IMDB.
Your data folder with json
files have to contain only directories with names that are exactly matching project dataset names. You can also upload a .tar
archive with a data folder inside.
Look at the examples for Directory and Archive below:
1) Directory 2) Archive
. .
├── DataFolder ├── Metadata.tar
|── dataset01 |── DataFolder
| ├── IMG_521.json |── dataset01
| ├── IMG_522.json | ├── IMG_521.json
| └── . . . | ├── IMG_522.json
└── dataset02 | └── . . .
├── IMG_4561.json └── dataset02
├── IMG_4562.json ├── IMG_4561.json
└── . . . ├── IMG_4562.json
└── . . .
Step 1: Add app to your team from Ecosystem if it is not there.
Step 2: Open context menu of your project -> Run App
-> Import Metadata
Step 1 Drag and drop you folder or tar
archive with data to Team Files.
Step 2 Right click on your data to open context menu and choose copy path
option.
Important notice: copy path of your project folder instead of dataset folder.
Step 3 Run application from project context menu and paste path to your data.
Step 4 After running the application, you will be redirected to the tasks page. Once application processing has finished, you can go to your project to see the results.