CaptureVisionModel Object

A CaptureVisionModel object defines how the library find Convolutional Neural Networks (CNN) model files that support the barcode decoding, character recognition and text line recognition. All the officially provided model files are shown in the image below.

capture-vision-model

Supported CaptureVisionModel files

Example

{
    "Name": "NumberLetterCharRecognition",
    "DirectoryPath": "D:\\CaptureVisionModel\\",
    "MaxModelInstances": 1,
    "ModelArgs": {
        "CharSet": {}
    }
}

Available Parameters

Name

The name of the CaptureVisionModel object. It will be used as a unique identifier for the model.

Parameter Summary
Type String
Remarks It must be the filename of the model file. For example, to use the `NumberLetterCharRecognition.data` model file, you need to set `Name` to “NumberLetterCharRecognition”.

DirectoryPath

The directory path of the model file.

Parameter Summary
Type String
Default Value ””

MaxModelInstances

The maximum number of instances of the model.

Parameter Summary
Type int
Range [1,256]
Default Value 1 (4 for single character model)

ModelArgs

Sets advanced arguments of the model.

Parameter Summary
Type JSON object
Remarks Currently, CharSet is supported