To avoid service disruption, update to a newer model (for example,
gemini-2.5-flash-lite). Learn more.
Replace the background of an image using Imagen
Stay organized with collections
Save and categorize content based on your preferences.
This page describes how to use Imagen to replace the background of an image using the Firebase AI Logic SDKs.
Background replacement is a type of mask-based editing (specifically inpainting). A mask is a digital overlay defining the specific area you want to edit.
How it works: You provide an original image and a corresponding masked image that defines a mask over the background — either using automatic background detection or providing the mask of the background yourself. You also provide a text prompt describing what you want to change. The model then generates and applies a new background.
For example, you can change the setting around a subject or object without affecting the foreground (for example, in a product image).
Jump to code for auto-detected background Jump to code for providing the background mask
Before you begin
If you haven't already, complete the
getting started guide, which
describes how to set up your Firebase project, connect your app to Firebase,
add the SDK, initialize the backend service for your chosen API provider, and
create an ImagenModel instance.
Models that support this capability
Imagen offers image editing through its capability
model:
imagen-3.0-capability-001
Note that for Imagen models, the global location is
not supported.
Replace background using automatic background detection
The following sample shows how to replace the background of an image — using automatic background detection. You provide the original image and a text prompt, and Imagen automatically detects and creates a mask of the background to modify the original image.
Swift
Image editing with Imagen models isn't supported for Swift. Check back later this year!
Kotlin
To replace the background using automatic background detection, specify
ImagenBackgroundMask. Use
editImage()
and set the editing config to use ImagenEditMode.INPAINT_INSERTION.
// Using this SDK to access Imagen models is a Preview release and requires opt-in
@OptIn(PublicPreviewAPI::class)
suspendfuncustomizeImage(){
// Initialize the Vertex AI Gemini API backend service
// Optionally specify the location to access the model (for example, `us-central1`)
valai=Firebase.ai(backend=GenerativeBackend.vertexAI(location="us-central1"))
// Create an `ImagenModel` instance with an Imagen "capability" model
valmodel=ai.imagenModel("imagen-3.0-capability-001")
// This example assumes 'originalImage' is a pre-loaded Bitmap.
// In a real app, this might come from the user's device or a URL.
valoriginalImage:Bitmap=TODO("Load your original image Bitmap here")
// Provide the prompt describing the new background.
valprompt="space background"
// Use the editImage API to replace the background.
// Pass the original image, the prompt, and an editing configuration.
valeditedImage=model.editImage(
referenceImages=listOf(
ImagenRawImage(originalImage.toImagenInlineImage()),
ImagenBackgroundMask(),// Use ImagenBackgroundMask() to auto-generate the mask.
),
prompt=prompt,
// Define the editing configuration for inpainting and background replacement.
config=ImagenEditingConfig(ImagenEditMode.INPAINT_INSERTION)
)
// Process the resulting 'editedImage' Bitmap, for example, by displaying it in an ImageView.
}
Java
To replace the background using automatic background detection, specify
ImagenBackgroundMask. Use
editImage()
and set the editing config to use ImagenEditMode.INPAINT_INSERTION.
// Initialize the Vertex AI Gemini API backend service
// Optionally specify the location to access the model (for example, `us-central1`)
// Create an `ImagenModel` instance with an Imagen "capability" model
ImagenModelimagenModel=FirebaseAI.getInstance(GenerativeBackend.vertexAI("us-central1"))
.imagenModel(
/* modelName */"imagen-3.0-capability-001");
ImagenModelFuturesmodel=ImagenModelFutures.from(imagenModel);
// This example assumes 'originalImage' is a pre-loaded Bitmap.
// In a real app, this might come from the user's device or a URL.
BitmaporiginalImage=null;// TODO("Load your image Bitmap here");
// Provide the prompt describing the new background.
Stringprompt="space background";
// Define the list of sources for the editImage call.
// This includes the original image and the auto-generated mask.
ImagenRawImagerawOriginalImage=
newImagenRawImage(ImagenInlineImageKt.toImagenInlineImage(originalImage));
// Use ImagenBackgroundMask() to auto-generate the mask.
ImagenBackgroundMaskrawMaskedImage=newImagenBackgroundMask();
ImagenEditingConfigconfig=newImagenEditingConfig();
// Use the editImage API to replace the background.
// Pass the original image, the auto-generated masked image, the prompt, and an editing configuration.
Futures.addCallback(model.editImage(Arrays.asList(rawOriginalImage,rawMaskedImage),prompt,config),
newFutureCallback<ImagenGenerationResponse>(){
@Override
publicvoidonSuccess(ImagenGenerationResponseresult){
if(result.getImages().isEmpty()){
Log.d("ImageEditor","No images generated");
}
BitmapeditedImage=((ImagenInlineImage)result.getImages().get(0)).asBitmap();
// Process and use the bitmap to display the image in your UI
}
@Override
publicvoidonFailure(Throwablet){
// ...
}
},Executors.newSingleThreadExecutor());
Web
Image editing with Imagen models isn't supported for Web apps. Check back later this year!
Dart
To replace the background using automatic background detection, specify
ImagenBackgroundMask. Use
editImage()
and set the editing config to use ImagenEditMode.inpaintInsertion.
import'dart:typed_data';
import'package:firebase_ai/firebase_ai.dart';
import'package:firebase_core/firebase_core.dart';
import'firebase_options.dart';
// Initialize FirebaseApp
awaitFirebase.initializeApp(
options:DefaultFirebaseOptions.currentPlatform,
);
// Initialize the Vertex AI Gemini API backend service
// Optionally specify a location to access the model (for example, `us-central1`)
finalai=FirebaseAI.vertexAI(location:'us-central1');
// Create an `ImagenModel` instance with an Imagen "capability" model
finalmodel=ai.imagenModel(model:'imagen-3.0-capability-001');
// This example assumes 'originalImage' is a pre-loaded Uint8List.
// In a real app, this might come from the user's device or a URL.
finalUint8ListoriginalImage=Uint8List(0);// TODO: Load your original image data here.
// Provide the prompt describing the new background.
finalprompt='space background';
try{
// Use the editImage API to replace the background.
// Pass the original image, the prompt, and an editing configuration.
finalresponse=awaitmodel.editImage(
sources:[
ImagenRawImage(originalImage),
ImagenBackgroundMask(),// Use ImagenBackgroundMask() to auto-generate the mask.
],
prompt:prompt,
// Define the editing configuration for inpainting and background replacement.
config:constImagenEditingConfig(
editMode:ImagenEditMode.inpaintInsertion,
),
);
// Process the result.
if(response.images.isNotEmpty){
finaleditedImage=response.images.first.bytes;
// Use the editedImage (a Uint8List) to display the image, save it, etc.
print('Image successfully generated!');
}else{
// Handle the case where no images were generated.
print('Error: No images were generated.');
}
}catch(e){
// Handle any potential errors during the API call.
print('An error occurred: $e');
}
Unity
Image editing with Imagen models isn't supported for Unity. Check back later this year!
Replace background using a provided mask
The following sample shows how to replace the background of an image — using a background mask defined in an image that you provide. You provide the original image, a text prompt, and the masked image.
Swift
Image editing with Imagen models isn't supported for Swift. Check back later this year!
Kotlin
To replace the background using a mask that you provide, specify
ImagenRawMask with the masked image. Use
editImage()
and set the editing config to use ImagenEditMode.INPAINT_INSERTION.
// Using this SDK to access Imagen models is a Preview release and requires opt-in
@OptIn(PublicPreviewAPI::class)
suspendfuncustomizeImage(){
// Initialize the Vertex AI Gemini API backend service
// Optionally specify the location to access the model (for example, `us-central1`)
valai=Firebase.ai(backend=GenerativeBackend.vertexAI(location="us-central1"))
// Create an `ImagenModel` instance with an Imagen "capability" model
valmodel=ai.imagenModel("imagen-3.0-capability-001")
// This example assumes 'originalImage' is a pre-loaded Bitmap.
// In a real app, this might come from the user's device or a URL.
valoriginalImage:Bitmap=TODO("Load your original image Bitmap here")
// This example assumes 'maskImage' is a pre-loaded Bitmap that contains the masked area.
// In a real app, this might come from the user's device or a URL.
valmaskImage:Bitmap=TODO("Load your masked image Bitmap here")
// Provide the prompt describing the new background.
valprompt="space background"
// Use the editImage API to replace the background.
// Pass the original image, the masked image, the prompt, and an editing configuration.
valeditedImage=model.editImage(
referenceImages=listOf(
ImagenRawImage(originalImage.toImagenInlineImage()),
ImagenRawMask(maskImage.toImagenInlineImage()),// Use ImagenRawMask() to provide your own masked image.
),
prompt=prompt,
// Define the editing configuration for inpainting and background replacement.
config=ImagenEditingConfig(ImagenEditMode.INPAINT_INSERTION)
)
// Process the resulting 'editedImage' Bitmap, for example, by displaying it in an ImageView.
}
Java
To replace the background using a mask that you provide, specify
ImagenRawMask with the masked image. Use
editImage()
and set the editing config to use ImagenEditMode.INPAINT_INSERTION.
// Initialize the Vertex AI Gemini API backend service
// Optionally specify the location to access the model (for example, `us-central1`)
// Create an `ImagenModel` instance with an Imagen "capability" model
ImagenModelimagenModel=FirebaseAI.getInstance(GenerativeBackend.vertexAI("us-central1"))
.imagenModel(
/* modelName */"imagen-3.0-capability-001");
ImagenModelFuturesmodel=ImagenModelFutures.from(imagenModel);
// This example assumes 'originalImage' is a pre-loaded Bitmap.
// In a real app, this might come from the user's device or a URL.
BitmaporiginalImage=null;// TODO("Load your original image Bitmap here");
// This example assumes 'maskImage' is a pre-loaded Bitmap that contains the masked area.
// In a real app, this might come from the user's device or a URL.
BitmapmaskImage=null;// TODO("Load your masked image Bitmap here");
// Provide the prompt describing the new background.
Stringprompt="space background";
// Define the list of source images for the editImage call.
ImagenRawImagerawOriginalImage=
newImagenRawImage(ImagenInlineImageKt.toImagenInlineImage(originalImage));
// Use ImagenRawMask() to provide your own masked image.
ImagenBackgroundMaskrawMaskedImage=
newImagenRawMask(ImagenInlineImageKt.toImagenInlineImage(maskImage));
ImagenEditingConfigconfig=newImagenEditingConfig();
// Use the editImage API to replace the background.
// Pass the original image, the masked image, the prompt, and an editing configuration.
Futures.addCallback(model.editImage(Arrays.asList(rawOriginalImage,rawMaskedImage),prompt,config),
newFutureCallback<ImagenGenerationResponse>(){
@Override
publicvoidonSuccess(ImagenGenerationResponseresult){
if(result.getImages().isEmpty()){
Log.d("ImageEditor","No images generated");
}
BitmapeditedImage=((ImagenInlineImage)result.getImages().get(0)).asBitmap();
// Process and use the bitmap to display the image in your UI
}
@Override
publicvoidonFailure(Throwablet){
// ...
}
},Executors.newSingleThreadExecutor());
Web
Image editing with Imagen models isn't supported for Web apps. Check back later this year!
Dart
To replace the background using a mask that you provide, specify
ImagenRawMask with the masked image. Use
editImage()
and set the editing config to use ImagenEditMode.INPAINT_INSERTION.
import'dart:typed_data';
import'package:firebase_ai/firebase_ai.dart';
import'package:firebase_core/firebase_core.dart';
import'firebase_options.dart';
// Initialize FirebaseApp
awaitFirebase.initializeApp(
options:DefaultFirebaseOptions.currentPlatform,
);
// Initialize the Vertex AI Gemini API backend service
// Optionally specify a location to access the model (for example, `us-central1`)
finalai=FirebaseAI.vertexAI(location:'us-central1');
// Create an `ImagenModel` instance with an Imagen "capability" model
finalmodel=ai.imagenModel(model:'imagen-3.0-capability-001');
// This example assumes 'originalImage' is a pre-loaded Uint8List.
// In a real app, this might come from the user's device or a URL.
finalUint8ListoriginalImage=Uint8List(0);// TODO: Load your original image data here.
// This example assumes 'maskImage' is a pre-loaded Uint8List that contains the masked area.
// In a real app, this might come from the user's device or a URL.
finalUint8ListmaskImage=Uint8List(0);// TODO: Load your masked image data here.
// Provide the prompt describing the new background.
finalprompt='space background';
try{
// Use the editImage API to replace the background.
// Pass the original image, the prompt, and an editing configuration.
finalresponse=awaitmodel.editImage(
sources:[
ImagenRawImage(originalImage),
ImagenRawMask(maskImage),// Use ImagenRawMask() to provide your own masked image.
],
prompt:prompt,
// Define the editing configuration for inpainting and background replacement.
config:constImagenEditingConfig(
editMode:ImagenEditMode.inpaintInsertion,
),
);
// Process the result.
if(response.images.isNotEmpty){
finaleditedImage=response.images.first.bytes;
// Use the editedImage (a Uint8List) to display the image, save it, etc.
print('Image successfully generated!');
}else{
// Handle the case where no images were generated.
print('Error: No images were generated.');
}
}catch(e){
// Handle any potential errors during the API call.
print('An error occurred: $e');
}
Unity
Image editing with Imagen models isn't supported for Unity. Check back later this year!
Best practices and limitations
We recommend dilating the mask when editing an image. This can help smooth
the borders of an edit and make it seem more convincing. Generally, a dilation
value of 1% or 2% is recommended (0.01 or 0.02).
Give feedback about your experience with Firebase AI Logic