Importing Annotations

Overview

You can import annotations while creating a Dataset in Visual Layer Profiler. The supported annotation types are:

• Image annotations: Images with class labels.
• Object annotations: Objects annotated with bounding boxes and corresponding labels.

Supported Annotation Formats

1. Parquet / CSV
2. JSON (COCO format)

Limitations ⚠️

• Annotations cannot be added after a Dataset has been created.
• The annotation file must be named as one of the following:
  • annotations.json
  • image_annotations.csv
  • object_annotations.csv
  • image_annotations.parquet
  • object_annotations.parquet

Parquet / CSV Annotations

Image Annotations

For full-image class labels, create an annotation file named image_annotations.parquet or image_annotations.csv. Place this file in the root directory of your uploaded folder, tar, or zip file.

Example Format:

import pandas as pd
df = pd.read_parquet('image_annotations.parquet')
df

• The filename column must contain relative paths.
• You may also include a caption column for textual metadata.
• Multiple labels can be stored as a list, e.g., ['t-shirt', 'SKU12345'].

Example for Multiple Labels:

Object Annotations

For object-level annotations, create an annotation file named object_annotations.parquet or object_annotations.csv. Each row represents a single object, with coordinates defining its bounding box.

Example Format:

import pandas as pd
df = pd.read_parquet('object_annotations.parquet')
df

• col_x and row_y are the top-left corner coordinates.
• width and height must be greater than zero.

JSON Annotations (COCO Format)

To use JSON annotations, ensure:

• The file is named annotations.json.
• It is placed in the root directory of your uploaded folder or S3 bucket.
• The file is present at the time of Dataset creation (annotations cannot be added later).

Example JSON Format:

{
    "images": [
        { "id": 1, "width": 640, "height": 480, "file_name": "image1.jpg" },
        { "id": 2, "width": 800, "height": 600, "file_name": "image2.jpg" }
    ],
    "categories": [
        { "id": 1, "name": "cat" },
        { "id": 2, "name": "dog" },
        { "id": 3, "name": "t-rex" }
    ],
    "annotations": [
        { "id": 1, "image_id": 1, "category_id": 1, "bbox": [100, 100, 200, 200] },
        { "id": 2, "image_id": 2, "category_id": 2, "bbox": [50, 50, 150, 150] },
        { "id": 3, "image_id": 1, "category_id": 3 },
        { "id": 4, "image_id": 2, "category_id": 3 }
    ]
}

• Bounding boxes (bbox) follow the format [col_x, row_y, width, height].
• Remove comments before uploading.

By following these guidelines, you can successfully import annotations into Visual Layer Profiler.