> For the complete documentation index, see [llms.txt](https://ai-tutorial.dionny.dev/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ai-tutorial.dionny.dev/exercises/exercise-1-object-detection.md).

# Exercise: Object Detection

## Google Cloud Vision API

1. Navigate to <https://cloud.google.com/vision>.&#x20;
2. Scroll down to the `Try the API` section.

![Try Cloud Vision API](/files/-Ml8ASp7G0uyxa6wfAcY)

1. Drag in an image. Here's one you can use:

![Sample Application Screen](/files/-Ml8ASp8PyT_4nbVi3Rg)

{% hint style="danger" %}
You'll need to solve a Captcha Challenge that is easy to miss (you may need to scroll down).
{% endhint %}

## Viewing the Results

When viewing the results, you will be able to see `Objects`, `Labels`, `Text`, and `Properties`. Click on the tabs below to learn more about each of the result types.

You can also see the raw JSON request and response payloads, which are useful when integrating with Google Cloud Vision API.

{% tabs %}
{% tab title="Objects" %}
The below results show us that Google Cloud Vision API was able to find two objects:

* `lighting` with `82%` confidence
* `couch` with `54%` confidence

![](/files/-Ml8ASp9oP9Sf7sgfa6a)

{% hint style="success" %}
**Discussion Topic:** How useful do you think this would be for testing purposes?
{% endhint %}
{% endtab %}

{% tab title="Labels" %}
According to Google Cloud Vision API, the following labels may be applied

![Suggested Labels Sorted by Confidence](/files/-Ml8ASpAH1dtdROK356F)

{% hint style="success" %}
**Discussion Topic:** How useful do you think this would be for testing purposes?
{% endhint %}
{% endtab %}

{% tab title="Text" %}
According to Google Cloud Vision API, the following text blocks were detected:

![](/files/-Ml8ASpB2Mo5MZ7pS2nM)

{% hint style="success" %}
**Discussion Topic:** How useful do you think this would be for testing purposes?
{% endhint %}
{% endtab %}

{% tab title="Properties" %}
Google Cloud Vision API also provides a few additional details about the image, such as the dominant colors:

![Dominant Colors and Crop Hints](/files/-Ml8ASpCPPCF1VVNZqcr)
{% endtab %}

{% tab title="Request/Response Payloads" %}
![Request/Response Payloads](/files/-Ml8ASpDOhDiVtDjOVOf)
{% endtab %}
{% endtabs %}

## Meta: Segment Anything

Try out Meta's SAM too!

<https://segment-anything.com/demo>