cv2/README.md

Image Decode Implementation with CV2

author: Kaiyuan Hu, Rentong Guo

Description

An image decode operator implementation with OpenCV.

Code Example

Load a image from path './src_dog.jpg'.

the src picture:

Write the pipeline in simplified style:

from towhee import pipe, ops, DataCollection


# decode image, in bgr channel order
p = (
  pipe.input('url')
  .map('url', 'image', ops.image_decode.cv2())
  .output('image')
)

# decode image, in rgb channel order
p2 = (
  pipe.input('url')
  .map('url', 'image', ops.image_decode.cv2('rgb'))
  .output('image')
)

# decode from path
DataCollection(p('./src_dog.jpg')).show()

# decode from bytes
with open('./src_dog.jpg', 'rb') as f:
  DataCollection(p2(f.read())).show()

Factory Constructor

Create the operator via the following factory method:

ops.image_decode.cv2()

mode: str

BGR or RGB, default is BGR

Interface

An image decode operator takes an image path as input. It decodes the image back to ndarray.

Parameters:

img: str

​ Local file path/http url/image bytes.

Returns: towhee.types.Image (a sub-class of numpy.ndarray)

​ The decoded image data as numpy.ndarray.