EasyObject - Analyzing Blobs¶

EasyObject is the blob analysis library of eVision. A blob is a grouping of neighboring pixels of the same gray level range. The pixels from a blob are connected: there exists a path between any couple of pixels from the blob that is entirely within the blob. This is also called connected-componenent analysis.

EasyObject can be used on BW8 ^.NET ^Python, BW16 ^.NET ^Python and C24 ^.NET ^Python images and its processing can be restricted to ROIs and to complex regions.

Process and Workflow¶

        ---
title: Basic workflow of EasyObject
---
flowchart LR
    Image["Image"]

    subgraph Encoder["Image encoder (EImageEncoder)"]
        direction TB
        Segment --> Blobs["Build blobs"]
    end

    Image --> Encoder

    Encoder --> CodedImage["Coded image (ECodedImage2)"]

    CodedImage -->|select| ObjectSelection["Object selection (EObjectSelection)"]

    ObjectSelection -->|sort| ObjectSelection

    ObjectSelection -->  Objects@{ shape: docs, label: "Array of blobs"}

It works by encoding an image into a coded image. The encoder ^.NET ^Python classifies the image into several layers with image segmenters and builds the blobs from the selected layers.
It can compute various features for each objects (for example its area, bounding box, etc.).
These features can be used to select, filter, or sort the blobs with EObjectSelection ^.NET ^Python.
The object template matcher ^.NET ^Python can be used to align and match the blobs to a reference image.
There are two kinds of blobs:
- objects that are built directly by the encoder from the segmentation of the image
- holes that are built from the objects.
- Both types of blobs have the same class, named ECodedElement ^.NET ^Python.