# LocalizationModes

Determines how to localize text lines. It consists of one or more modes, each mode representing a different localization process.

**Remarks**

- Introduced in Dynamsoft Capture Vision version 3.2.1000.

## JSON Structure

**Location in template:**

```
LabelRecognizerTaskSettingOptions[i]
    └── SectionArray[j] (Section object)
        └── StageArray[k] (Stage object where Stage = "SST_LOCALIZE_TEXT_LINES")
            └── LocalizationModes
```

**Parent object:** [LocalizeTextLinesStage](https://www.dynamsoft.com/capture-vision/docs/core/parameters/reference/label-recognizer-task-settings/stage-localize-text-lines.md)

**Example:**

```json
"LocalizationModes":
[
    {
        "Mode": "LM_NEURAL_NETWORK",
        "ModelNameArray": ["MRZLocalization"]
    },
    {
        "Mode": "LM_GENERAL"
    }
]
```

> [!NOTE]
> - This snippet shows only the `LocalizationModes` parameter.
> - To use it, embed this parameter within a [LocalizeTextLinesStage](https://www.dynamsoft.com/capture-vision/docs/core/parameters/reference/label-recognizer-task-settings/stage-localize-text-lines.md) object.
> - For the complete JSON structure, see:
>   - [Full JSON Structure](https://www.dynamsoft.com/capture-vision/docs/core/parameters/file/index.md)
>   - [Minimal Valid JSON](https://www.dynamsoft.com/capture-vision/docs/core/parameters/file/index.md#minimal-valid-json)

## Parameter Details

Parameter `LocalizationModes` consist of a group of localization mode objects. Each localization mode object includes a candidate mode and a series of auxiliary mode arguments. The structure of the localization mode object is shown as follow:

### Mode Arguments

<table style = "text-align:left">
    <thead>
        <tr>
            <th nowrap="nowrap">Mode Argument Name</th>
            <th nowrap="nowrap">Mode Argument Summary</th>
        </tr>
    </thead>
    <tr>
        <td rowspan = "3" style="vertical-align:text-top">Mode<br>(Required)</td>
        <td><b>Description</b><br>Any one in Candidate Mode List as string
        </td>
    </tr>
    <tr>
        <td><b>Type</b><br><i>String</i>
        </td>
    </tr>
    <tr>
        <td><b>Candidate Mode List</b><br>
            LM_NEURAL_NETWORK<br>
            LM_GENERAL<br>
        </td>
    </tr>
    <tr>
        <td rowspan = "6" style="vertical-align:text-top" id="modelnamearray">ModelNameArray<br>(Optional)</td>
        <td><b>Description</b><br>Sets the names of a text localization model.
        </td>
    </tr>
    <tr>
        <td><b>Type</b><br><i>String Array</i>
        </td>
    </tr>
    <tr>
        <td><b>Range</b><br>Each element is the name of a model. Candidate values: "MRZLocalization".
        </td>
    </tr>
    <tr>
        <td><b>Default Value</b><br>null
        </td>
    </tr>
    <tr>
        <td><b>Valid For</b><br>LM_NEURAL_NETWORK.
        </td>
    </tr>
    <tr>
        <td><b>Remarks</b><br>When set to null, model "MRZLocalization" will be used by default.
        </td>
    </tr>
</table>

### Default Settings

If the `LocalizationModes` is not configured in your template file, the following settings will be used as the default settings.

```json
{
    "LocalizationModes":
    [
        {
            "Mode": "LM_GENERAL"
        }
    ]
}
```

## Candidate Modes Descriptions

### LM_GENERAL

Localizes text lines using general image processing algorithms.

### LM_NEURAL_NETWORK

Localizes text lines by utilizing a neural network model. This mode has the following arguments for further cuztomization.

**Available Mode Arguments:**

- ModelNameArray