Image | Documentation

The Image component is an operator component that allows users to manipulate image files. It can carry out the following tasks:

Concat
Crop
Resize
Draw Classification
Draw Detection
Draw Keypoint
Draw OCR
Draw Instance Segmentation
Draw Semantic Segmentation

#Release Stage

Alpha

#Configuration

The component definition and tasks are defined in the definition.yaml and tasks.yaml files respectively.

#Supported Tasks

#Concat

Concatenate images horizontally or vertically. All images must have the same width and height. If grid-width and grid-height are not provided, the images will be concatenated to a square grid.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_CONCAT`
Images (required)	`images`	array	Input images.
Grid Width	`grid-width`	integer	Grid width. If `grid-width` is provided, `grid-height` will be ignored.
Grid Height	`grid-height`	integer	Grid height. If `grid-height` is provided, `grid-width` will be ignored.
Padding	`padding`	integer	Padding between images. If `padding` is provided, it will be applied to all four sides of the image.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Crop

Crop image to the specified size.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_CROP`
Image (required)	`image`	string	Input image.
Corner Radius	`corner-radius`	integer	Radius from the corner of the image to crop the image. If `corner-radius` is provided, `circle-radius` will be ignored. If the `corner-radius` is larger than half of min(width, height), the min(width, height) will be applied to `corner-radius` .
Circle Radius	`circle-radius`	integer	Radius from the center of the circle to crop the image. If `circle-radius` is provided, `corner-radius` will be ignored. If the `circle-radius` is larger than half of min(width, height), the min(width, height) will be applied to `circle-radius`.
Top Offset	`top-offset`	integer	Top offset of the crop.
Right Offset	`right-offset`	integer	Right offset of the crop.
Bottom Offset	`bottom-offset`	integer	Bottom offset of the crop.
Left Offset	`left-offset`	integer	Left offset of the crop.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Resize

Resize image to the specified size.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_RESIZE`
Image (required)	`image`	string	Input image.
Ratio	`ratio`	number	Resize ratio (e.g., 0.5 for half size, 2 for double size) to resize the image. If ratio is provided, width and height will be ignored.
Width	`width`	integer	Width of the output image.
Height	`height`	integer	Height of the output image.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Draw Classification

Draw classification result on the image.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_DRAW_CLASSIFICATION`
Category (required)	`category`	string	The predicted category of the input.
Image (required)	`image`	string	Input image.
Show Score	`show-score`	boolean	Show model confidence score on each instance.
Score (required)	`score`	number	The confidence score of the predicted category of the input.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Draw Detection

Draw detection result on the image.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_DRAW_DETECTION`
Image (required)	`image`	string	Input image.
Objects (required)	`objects`	array	A list of detected objects.
Show Score	`show-score`	boolean	Show model confidence score on each instance.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Draw Keypoint

Draw keypoint result on the image.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_DRAW_KEYPOINT`
Image (required)	`image`	string	Input image.
Objects (required)	`objects`	array	A list of keypoint objects, a keypoint object includes all the pre-defined keypoints of a detected object.
Show Score	`show-score`	boolean	Show model confidence score on each instance.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Draw OCR

Draw OCR result on the image.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_DRAW_OCR`
Image (required)	`image`	string	Input image.
Objects (required)	`objects`	array	A list of detected bounding boxes.
Show Score	`show-score`	boolean	Show model confidence score on each instance.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Draw Instance Segmentation

Draw instance segmentation result on the image.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_DRAW_INSTANCE_SEGMENTATION`
Image (required)	`image`	string	Input image.
Objects (required)	`objects`	array	A list of detected instance bounding boxes.
Show Score	`show-score`	boolean	Show model confidence score on each instance.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.

#Draw Semantic Segmentation

Draw semantic segmentation result on the image.

Input	Field ID	Type	Description
Task ID (required)	`task`	string	`TASK_DRAW_SEMANTIC_SEGMENTATION`
Image (required)	`image`	string	Input image.
Stuffs (required)	`stuffs`	array	A list of RLE binary masks.

Output	Field ID	Type	Description
Image	`image`	image/png	Output image.