# Getting Started

## Documentation and Samples

Technical information: [Celantur SDK Library](https://www.celantur.com/sdk/doc/current/)

 Samples: [SDKExample on GitHub](https://github.com/celantur/SDKExample).

## Get a Demo 

{% hint style="info" %}
[Reach out to our team](https://www.celantur.com/contact/) to request a demo version.
{% endhint %}

## Anonymizing Panorama Images

Settings:

* ROI: 0.0, 0.2 x 1.0, 0.6
* Tiling: 5 x 1
* Tile Overlap 0.1 x 0.0

{% embed url="<https://youtu.be/EGf5Noyb6rk>" %}
Real-time blurring of a dashcam live stream with Celantur Edge SDK on an Nvidia Jetson.
{% endembed %}

## Features

<table><thead><tr><th width="239.33331298828125">Feature</th><th width="119.99993896484375">Category</th><th width="298.33343505859375">Description</th><th>Available since</th></tr></thead><tbody><tr><td>Bounding box blurring</td><td>Blurring</td><td>Blur license plates (rectangular shape) and faces (oval shape).</td><td>v1.0</td></tr><tr><td>Segmentation blurring</td><td>Blurring</td><td>Blur segmentation masks</td><td>v1.1</td></tr><tr><td>License plate and face</td><td>Detection</td><td> </td><td>v1.0</td></tr><tr><td>Vehicle and person</td><td>Detection</td><td> </td><td>v1.1</td></tr><tr><td>Setting detection thresholds</td><td>Detection</td><td>User can set detection threshold</td><td>v1.0</td></tr><tr><td>Custom tiling</td><td>Detection</td><td>Control how many tiles an image is split into before inference.</td><td>v1.0</td></tr><tr><td>Image blurring</td><td>Image processing</td><td>Blurred image as 3-channel matrices.</td><td>v1.0</td></tr><tr><td>Encode and decode JPEG images</td><td>Image processing</td><td>JPEG images can be loaded and blurred images can be saved as JPEG images.</td><td>v1.0</td></tr><tr><td>Conserve JPEG metdata</td><td>Image processing</td><td>JPEG metadata (XMP, EXIF) are copied over in the anonymised image.</td><td>v1.0</td></tr><tr><td>Detection as structured data</td><td>Detection</td><td>Detection metadata (type and position of bounding box) can be retrieved.</td><td>v1.0</td></tr><tr><td>TensorRT GPU inference</td><td>Inference</td><td>Inference on NVIDIA GPU.</td><td>v1.1</td></tr><tr><td>ONNX CPU inference</td><td>Inference</td><td>Inference on Intel and AMD processors.</td><td>v1.0</td></tr><tr><td>OpenVINO CPU inference</td><td>Inference</td><td>Inference on Intel (and AMD) processors.</td><td> </td></tr><tr><td>Configure multithreading</td><td>Inference</td><td>Set number of threads for CPU processing.</td><td>v1.0</td></tr><tr><td>Object tracking</td><td>Video processing</td><td></td><td> v1.4</td></tr></tbody></table>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://doc.celantur.com/sdk/getting-started.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.