How to use fiftyone for exploring the instance segmentation of custom coco data?

Question

How to use fiftyone for exploring the instance segmentation of custom coco data? It has documentation for coco dataset but I couldn't find any resource for custom coco dataset.

Answer 1

You can find documentation for importing custom COCODetectionDatasets here.

In short, as long as your data follows the expected COCO format:

{
    "info": {...},
    "licenses": [
        {
            "id": 1,
            "name": "Attribution-NonCommercial-ShareAlike License",
            "url": "http://creativecommons.org/licenses/by-nc-sa/2.0/",
        },
        ...
    ],
    "categories": [
        ...
        {
            "id": 2,
            "name": "cat",
            "supercategory": "animal",
            "keypoints": ["nose", "head", ...],
            "skeleton": [[12, 14], [14, 16], ...]
        },
        ...
    ],
    "images": [
        {
            "id": 1,
            "license": 1,
            "file_name": "<filename0>.<ext>",
            "height": 480,
            "width": 640,
            "date_captured": null
        },
        ...
    ],
    "annotations": [
        {
            "id": 1,
            "image_id": 1,
            "category_id": 2,
            "bbox": [260, 177, 231, 199],
            "segmentation": [...],
            "keypoints": [224, 226, 2, ...],
            "num_keypoints": 10,
            "score": 0.95,
            "area": 45969,
            "iscrowd": 0
        },
        ...
    ]
}

The segmentation field format is defined here.

Then you can load it into FiftyOne with the following Python code:

import fiftyone as fo

name = "my-dataset"
dataset_dir = "/path/to/coco-detection-dataset"

# Create the dataset
dataset = fo.Dataset.from_dir(
    dataset_dir=dataset_dir,
    dataset_type=fo.types.COCODetectionDataset,
    name=name,
)

# View summary info about the dataset
print(dataset)

# Print the first few samples in the dataset
print(dataset.head())

Answer 2

It can be done by using COCODetectionDatasetImporter Class and set label_types=["detections","segmentations"] for seeing mask annotated images