> ## Documentation Index
> Fetch the complete documentation index at: https://notes.kodekloud.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Demo Image Masking

> This tutorial teaches how to use DALL·E 2’s image masking feature for editing pictures by defining masked regions and providing text prompts.

In this tutorial, you’ll learn how to use DALL·E 2’s image masking feature to edit an existing picture by defining a masked region and supplying a text prompt. This approach is perfect for seamlessly replacing objects, changing backgrounds, or adding new elements to your images.

## Prerequisites

* Python 3.7+
* The `openai` Python package (`pip install openai`)
* An OpenAI API key

<Callout icon="lightbulb" color="#1CB2FE">
  Make sure you’ve exported your API key as an environment variable:

  ```bash theme={null}
  export OPENAI_API_KEY="your_api_key_here"
  ```
</Callout>

## 1. Install and Authenticate

Begin by installing the client library and initializing your API key in Python:

```python theme={null}
import openai
import os
from IPython.display import Image

# Load API key from environment
openai.api_key = os.getenv("OPENAI_API_KEY")
```

## 2. Prepare Your Images

You need three files:

1. **Original image** (with the object you want to replace)
2. **Base image** (same image after removing the object)
3. **Mask image** (a transparent PNG that marks the area to edit)

```python theme={null}
display(Image(filename='images/dog_table.png'))        # Original image with dog
display(Image(filename='images/table.png'))            # Image with dog removed
display(Image(filename='images/table-masked.png'))     # Transparent mask placeholder
```

## 3. Perform the Edit

Use the `create_edit` endpoint to fill the masked area based on your prompt:

```python theme={null}
response = openai.Image.create_edit(
    image=open('images/table.png', 'rb'),
    mask=open('images/table-masked.png', 'rb'),
    prompt='A cat sitting on a dining table chair waiting for food',
    n=1,
    size='512x512'
)
edited_url = response['data'][0]['url']
display(Image(url=edited_url))
```

## 4. Generate Multiple Variations

If you’d like several options, increase `n` and iterate through the results:

```python theme={null}
response = openai.Image.create_edit(
    image=open('images/table.png', 'rb'),
    mask=open('images/table-masked.png', 'rb'),
    prompt='A cat sitting on a dining table chair waiting for food',
    n=3,
    size='512x512'
)
for result in response['data']:
    display(Image(url=result['url']))
```

<Callout icon="triangle-alert" color="#FF6B6B">
  Requesting a large number of edits or very high resolutions may incur higher usage costs. Monitor your [API usage dashboard](https://platform.openai.com/account/usage).
</Callout>

## 5. Parameter Reference

| Parameter | Type    | Description                                                   |
| --------- | ------- | ------------------------------------------------------------- |
| `image`   | File    | The base image without the masked object (PNG or JPEG).       |
| `mask`    | File    | A transparent PNG marking the area to edit (white = mask).    |
| `prompt`  | String  | Textual description of what should appear in the masked area. |
| `n`       | Integer | Number of edits to generate (1–10).                           |
| `size`    | String  | Output resolution: `256x256`, `512x512`, or `1024x1024`.      |

## 6. Generate Image Variations

Beyond masking, DALL·E 2 can also create variations of a single image without any prompt:

```python theme={null}
response = openai.Image.create_variation(
    image=open('images/table.png', 'rb'),
    n=4,
    size='512x512'
)
for img in response['data']:
    display(Image(url=img['url']))
```

For full details on both endpoints, see the [OpenAI Image API Reference](https://platform.openai.com/docs/api-reference/images).

<CardGroup>
  <Card title="Watch Video" icon="video" cta="Learn more" href="https://learn.kodekloud.com/user/courses/mastering-generative-ai-with-openai/module/ad3893f7-fba6-4142-a575-422006496a97/lesson/ff7f071e-87be-41fd-8f4c-0d3fa959250a" />

  <Card title="Practice Lab" icon="installation" cta="Learn more" href="https://learn.kodekloud.com/user/courses/mastering-generative-ai-with-openai/module/ad3893f7-fba6-4142-a575-422006496a97/lesson/0dfd8ad8-78ca-4419-9380-653e4f34b845" />
</CardGroup>