bounded-box.md

id	title
bounded-box	Bounded Box

Bounded box

A bounded box is an infinitely customizable image annotation component that can be used to select and tag areas within an image. It supports selection using specific points (landmarking) or drawing rectangular areas (bounding boxes). It can be used to create datasets for machine learning models or to annotate images for other purposes.

Properties

Property	Description	Expected value
Image URL	The URL or image data to show it on the component.	Get the image URL dynamically from database: {{queries.queryname.data[0].url}} or use image's base64 data
Default value	The data that will load the default bounded boxes over the image when the app is loaded.	Array of objects. Check the Default value data properties
Selector	The bounded box support selection using rectangle or point.	Click Fx to set the value RECTANGLE or POINT
List of labels	The list of label that will be displayed in the dropdown while selection in the bounded-box.	Labels in array format: {{['Tree', 'Car', 'Stree light']}}

Default value

Provide the data that will load the default bounding boxes over the image when the app is loaded. The data is expected to be an array of objects format.

Property	Description	Expected value
type	Sets the type of the bounded box.	RECTANGLE or POINT
width	Sets the width of the bounded box in pixels.	Numeric value. If the type value is POINT, set it to 0
height	Sets the height of the bounded box in pixels.	Numeric value. If the type value is POINT, set it to 0
x	Sets the x-coordinate(horizontal) position of the bounded box in the image.	Numerical value ex: 41
y	Sets the y-coordinate(vertical) position of the bounded box in the image.	Numerical value ex: 22
text	Sets the text value of the bounded box.	It should be one of the labels provided in the List of labels property

Example of default values:

[
    {
        type: 'RECTANGLE',
        width: 40,
        height: 24,
        x: 41,
        y: 12,
        text: 'Tree'
    },
    {
        type: 'POINT',
        width: 0,
        height: 0,
        x: 10.28,
        y: 81.14,
        text: 'Car'
    }
]

Events

Events are actions that can be triggered programmatically when the user interacts with the component. Click on the component handle to open its properties on the right. Go to the Events accordion and click on + Add handler.

Event	Description
On change	Triggered when the label from the dropdown in the selector is changed in the bounded box.

:::info Check Action Reference docs to get the detailed information about all the Actions. :::

General

Tooltip

A Tooltip is often used to specify the extra information when the user hovers the mouse pointer over the component. Once a value is set for Tooltip, hovering over the element will display the specified string as the tooltip text.

Layout

Show on desktop

Use this toggle to show or hide the component in the desktop view. You can dynamically configure the value by clicking on Fx and entering a logical expression that results in either true or false. Alternatively, you can directly set the values to {{true}} or {{false}}.

Show on mobile

Use this toggle to show or hide the component in the mobile view. You can dynamically configure the value by clicking on Fx and entering a logical expression that results in either true or false. Alternatively, you can directly set the values to {{true}} or {{false}}.

Styles

Style	Description	Expected value
Visibility	Toggle on or off to control the visibility of the component when the app is loaded	{{true}} or {{false}}, By default, it's set to {{true}}
Disable	Toggle on to disable the component.	{{true}} or {{false}}, By default, it's set to {{false}}
Box shadow	Sets the add shadow effects around a component's frame. You can specify the horizontal and vertical offsets(through X and Y sliders), blur and spread radius, and color of the shadow.	Values that represent x,y, blur, spread and color. Ex: 9px 11px 5px 5px #00000040

Exposed variables

variable	Description
annotations	This variable is an array of objects, where each object represents an annotation added to an image. The object contains the following keys: type, x, y, width, height, text, and id
annotations.type	There are two types of annotations: RECTANGLE and POINT
annotations.x	coordinates on the x axis
annotations.y	coordinates on the y axis
annotations.width	width of the annotation
annotations.height	height of the annotation
annotations.text	label selected for the annotation
annotations.id	unique ID of the annotation (system generated)

The values can be accessed dynamically using {{components.boundedbox1.annotations[0].text}} or {{components.boundedbox1.annotations[1].width}}

Component specific actions (CSA)

There are currently no CSA (Component-Specific Actions) implemented to regulate or control the bounding box component.