This extension provides capabilities for standard and advanced image processing and analytics, such as:

Image Data Ingestion

• Files (png, jpg, bmp...)
• Webcam
• IPCamera (experimental!)

Image Data Processing

Operators to do standard transformations like translating, rotating or croping of images.

Advanced Image Analytics

• Detect Objects in images using deep neural networks like YOLO
• Extract Text from images via Optical Character Recognition (OCR)
• Align an image to a template

This extension is based on OpenCV for most of it's operations.

Change Log

Changes in 0.5.0

• Added a new Operator Extract Image Information, which gives you the information like width, height, channels and image types. 
• Added a new Operator Add Regions, which adds regions (i.e. rectangles) to an image) 
• Added an option to Read Image to read the images of a PPTX file.
• Read Image has now a selector for the image type if you read PDFs.
• Renamed all occurrences of "subimage" to region
• Added a setting draw_regions to Write Image, which allows to also write the regions to the image
• The resulting tables from image to table now have their respective channel as an annotation. 
• Fixed a bug that made it impossible to retrieve images with bounding boxes.

Changes in 0.4.0

• Changed the storage format of larger images to make it compatible with RM 10.3. This can create the issue that larger images which are written in 10.2 cannot be loaded in 10.3.
• Made the max pixels in rendering a configurable. Its set to 800 by default
• Read Image Model is now actually working for DARKNET *
• Moving to openpnp's opencv packages instead of bytedeco. This allows this extension to run on more hardware. Most importantly ARM.
• Deprecated all model operators (Load Model, Load Model from Zoo, Detect Objects).

Version 0.2.2

• Added a new Operator Generate Pixel Aggregation of two images to aggregate on a pixel to pixel level.
• Added a new Operator Detect Lines to detect lines in Images