BoundingBox Annotations on Images

from aperturedb import Connector
from aperturedb import NotebookHelpers as nh

# ApertureDB Server Info for establishing connection
db_host  = "aperturedb.local"  # assuming local installation as provided
user     = "admin"             # requires authentication
password = "admin"             # use the password provided for the instance

db = Connector.Connector(db_host, user=user, password=password)

Introduce annotation on any image and then fetch just the pixel area within those coordinates

ApertureDB query language supports various regions of interest like BoundingBox, Polygon, to be associated with any image or frame.

query = [{
    "FindImage": {
        "_ref": 1,                        
        "constraints": {
            "id": ["==", 567989]
        },
        "blobs": False             # We don't need to fetch the actual pixels in this case
    }
}, {
    "AddBoundingBox": {
        "image_ref": 1,
        "rectangle": {
            "x": 40,               # pixels offset from the left edge
            "y": 50,               # pixels offset from the top edge
            "width": 450,          # box width in pixels
            "height": 400          # box height in pixels
        },
        "label": "Animal",
        "properties": {
            "name": "horse",
            "annotation_source": "manual"
        },
        "if_not_found": {         # Only add the bounding box if it doesn't already exist
            # note: These checks will operate faster if properties 'name' and 'annotation_source' are indexed
            "name": ["==", "horse"], 
            "annotation_source": ["==", "manual"]
        }
    }
}]

response, blobs = db.query(query)

db.print_last_response()

[
    {
        "FindImage": {
            "returned": 0,
            "status": 0
        }
    },
    {
        "AddBoundingBox": {
            "status": 0
        }
    }
]

query = [ {
    "FindImage": {
        "_ref": 1,
        "constraints": {
            "id": ["==", 567989]
        },
        "blobs": False             # Don't retrieve pixel data from the image...
    }
}, {
    "FindBoundingBox": {
        "is_connected_to": {
            "ref": 1
        },
        "with_label": "Animal",
        "blobs": True              # Instead, retrieve pixel data from the bounding box
    }
}]

res, blobs = db.query(query)

# This will display the image cropped to the bounding box
nh.display(blobs)

Find an image by its label

query = [ {
    "FindBoundingBox": {
        "_ref": 1,
        "with_label": "Animal",
        "coordinates": True,
        "labels": True,
        "blobs": False
    }}, {
    "FindImage": {
        "is_connected_to": {
            "ref": 1
        },
        "blobs": True
    }
} ]

res, blobs = db.query(query)

db.print_last_response()
nh.display(blobs)

[
    {
        "FindBoundingBox": {
            "entities": [
                {
                    "_coordinates": {
                        "height": 400,
                        "width": 450,
                        "x": 40,
                        "y": 50
                    },
                    "_label": "Animal"
                }
            ],
            "returned": 1,
            "status": 0
        }
    },
    {
        "FindImage": {
            "blobs_start": 0,
            "returned": 1,
            "status": 0
        }
    }
]

Delete the bounding box.

query = [ {
    "DeleteBoundingBox": {
        "constraints": {
            "annotation_source": ["==", "manual"]
        }
    }
}]

res, blobs = db.query(query)

db.print_last_response()

[
    {
        "DeleteBoundingBox": {
            "count": 1,
            "status": 0
        }
    }
]