Can't see annotation in CVAT after view.annotate(...)

[shaojun]

Hi,
I'm trying to download a very small subset of Open Images data of class Bicycle, and then upload the images and annotations to CVAT.

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Linux Ubuntu 20.04.3 LTS, x64
• FiftyOne installed from (pip or source): pip
• FiftyOne version (run fiftyone --version): FiftyOne v0.14.4, Voxel51, Inc.
• Python version: Python 3.8.10 (default, Nov 26 2021, 20:14:08)

Commands to reproduce

this is what I've done:

dataset = foz.load_zoo_dataset("open-images-v6", split="validation", label_types=["detections"], classes=["Bicycle"], max_samples=20, seed=51, shuffle=True, dataset_name="open-images-bicycle-mini-set-upload-to-cvat",)
session = fo.launch_app(dataset)
#can see the preview in below picture 1.
view = dataset.view()
anno_key="cvat_existing_field"
view.annotate(anno_key, label_field="ground_truth",label_type="detections",classes=["Bicycle","Person"])
#can see below info indicates the upload is successful, and open CVAT can see a new task is created with upload images, in picture 2
Uploading samples to CVAT...
Computing image metadata...
 100% |█████████████████████████████████████████████████████████████████| 20/20 [78.8ms elapsed, 0s remaining, 253.7 samples/s] 
Upload complete

picture 1:

picture 2:

Describe the problem

In CVAT, all upload images can be seen, but none of it has the annotation rectangles.

The tutorial works good.

What areas of FiftyOne does this bug affect?

• App: FiftyOne application issue
• Core: Core fiftyone Python library issue
• Server: Fiftyone server issue

Willingness to contribute

The FiftyOne Community encourages bug fix contributions. Would you or another
member of your organization be willing to contribute a fix for this bug to the
FiftyOne codebase?

• Yes. I can contribute a fix for this bug independently.
• Yes. I would be willing to contribute a fix for this bug with guidance
  from the FiftyOne community.
• No. I cannot contribute a bug fix at this time.

[brimoor]

In your dataset, your existing labels are in a field called "detections", not "ground_truth":

view.annotate(anno_key, label_field="detections", label_type="detections", classes=["Bicycle", "Person"])

Your current annotate() call is requesting bboxes in a new field called "ground_truth".

[shaojun]

thanks @brimoor
could you explain a bit more for label_field in view.annotate(...)

[brimoor]

These pages contain tons of info:

• https://voxel51.com/docs/fiftyone/integrations/cvat.html#requesting-annotations
• https://voxel51.com/docs/fiftyone/integrations/cvat.html#adding-new-label-fields
• https://voxel51.com/docs/fiftyone/integrations/cvat.html#editing-existing-labels

[fshaker]

thanks @brimoor could you explain a bit more for label_field in view.annotate(...)

Hi @shaojun , did you fix the problem? If yes how did you do it. I am having similar problem.