feat: imagen 3 inpainting and outpainting samples

generative_ai/image_generation/imagen3_editing/edit_image_3_inpainting_insert_mask.py

@@ -0,0 +1,80 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Google Cloud Vertex AI sample for editing an image using a mask file with Imagen 3.
+    Inpainting can insert the object designated by the prompt into the masked
+    area.
+"""
+import os
+
+from vertexai.preview import vision_models
+
+PROJECT_ID = os.getenv("GOOGLE_CLOUD_PROJECT")
+
+
+def edit_image_3_inpainting_insert_mask(
+    input_file: str,
+    mask_file: str,
+    output_file: str,
+    prompt: str,
+) -> vision_models.ImageGenerationResponse:
+    # [START generativeaionvertexai_imagen_3_edit_image_inpainting_insert_mask]
+
+    import vertexai
+    from vertexai.preview.vision_models import Image, ImageGenerationModel, MaskReferenceImage, RawReferenceImage
+
+    # TODO(developer): Update and un-comment below lines
+    # PROJECT_ID = "your-project-id"
+    # input_file = "input-image.png"
+    # mask_file = "mask-image.png"
+    # output_file = "output-image.png"
+    # prompt = "" # The text prompt describing what you want to see inserted.
+
+    vertexai.init(project=PROJECT_ID, location="us-central1")
+
+    model = ImageGenerationModel.from_pretrained("imagen-3.0-capability-001")
+    base_img = Image.load_from_file(location=input_file)
+    mask_img = Image.load_from_file(location=mask_file)
+    base_ref_image = RawReferenceImage(image=base_img, reference_id=0)
+    mask_ref_image = MaskReferenceImage(
+        reference_id=1, image=mask_img, mask_mode="user_provided"
+    )
+
+    images = model.edit_image(
+        reference_images=[base_ref_image, mask_ref_image],
+        prompt=prompt,
+        edit_mode="inpainting-insert",
+    )
+
+    images[0].save(location=output_file, include_generation_parameters=False)
+
+    # Optional. View the edited image in a notebook.
+    # images[0].show()
+
+    print(f"Created output image using {len(images[0]._image_bytes)} bytes")
+    # Example response:
+    # Created output image using 1400814 bytes
+
+    # [END generativeaionvertexai_imagen_3_edit_image_inpainting_insert_mask]
+
+    return images
+
+
+if __name__ == "__main__":
+    edit_image_3_inpainting_insert_mask(
+        input_file="../test_resources/woman.png",
+        mask_file="../test_resources/woman_inpainting_insert_mask.png",
+        output_file="../test_resources/woman_with_hat.png",
+        prompt="red hat",
+    )

generative_ai/image_generation/imagen3_editing/edit_image_3_inpainting_insert_mask_mode.py

@@ -0,0 +1,80 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Google Cloud Vertex AI sample for editing an image using a mask mode. The
+    mask mode is used to automatically select the background, foreground (i.e.,
+    the primary subject of the image), or an object based on segmentation class.
+    Inpainting can insert the object or background designated by the prompt.
+"""
+import os
+
+from vertexai.preview import vision_models
+
+PROJECT_ID = os.getenv("GOOGLE_CLOUD_PROJECT")
+
+
+def edit_image_3_inpainting_insert_mask_mode(
+    input_file: str,
+    mask_mode: str,
+    output_file: str,
+    prompt: str,
+) -> vision_models.ImageGenerationResponse:
+    # [START generativeaionvertexai_imagen_edit_image_3_inpainting_insert_mask_mode]
+
+    import vertexai
+    from vertexai.preview.vision_models import Image, ImageGenerationModel, MaskReferenceImage, RawReferenceImage
+
+    # TODO(developer): Update and un-comment below lines
+    # PROJECT_ID = "your-project-id"
+    # input_file = "input-image.png"
+    # mask_mode = "background" # 'background', 'foreground', or 'semantic'
+    # output_file = "output-image.png"
+    # prompt = "" # The text prompt describing what you want to see inserted.
+
+    vertexai.init(project=PROJECT_ID, location="us-central1")
+
+    model = ImageGenerationModel.from_pretrained("imagen-3.0-capability-001")
+    base_img = Image.load_from_file(location=input_file)
+    base_ref_image = RawReferenceImage(image=base_img, reference_id=0)
+    mask_ref_image = MaskReferenceImage(
+        reference_id=1, image=None, mask_mode=mask_mode
+    )
+
+    images = model.edit_image(
+        reference_images=[base_ref_image, mask_ref_image],
+        prompt=prompt,
+        edit_mode="inpainting-insert",
+    )
+
+    images[0].save(location=output_file, include_generation_parameters=False)
+
+    # Optional. View the edited image in a notebook.
+    # images[0].show()
+
+    print(f"Created output image using {len(images[0]._image_bytes)} bytes")
+    # Example response:
+    # Created output image using 1234567 bytes
+
+    # [END generativeaionvertexai_imagen_edit_image_3_inpainting_insert_mask_mode]
+
+    return images
+
+
+if __name__ == "__main__":
+    edit_image_3_inpainting_insert_mask_mode(
+        input_file="../test_resources/woman.png",
+        mask_mode="background",
+        output_file="../test_resources/woman_at_beach.png",
+        prompt="beach",
+    )

generative_ai/image_generation/imagen3_editing/edit_image_3_inpainting_insert_mask_mode_test.py

@@ -0,0 +1,41 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+
+import backoff
+
+import edit_image_3_inpainting_insert_mask_mode
+
+from google.api_core.exceptions import ResourceExhausted
+
+_RESOURCES = os.path.join(os.path.dirname(__file__), "../test_resources")
+_INPUT_FILE = os.path.join(_RESOURCES, "woman.png")
+_MASK_MODE = "background"
+_OUTPUT_FILE = os.path.join(_RESOURCES, "woman_at_beach.png")
+_PROMPT = "beach"
+
+
+@backoff.on_exception(backoff.expo, ResourceExhausted, max_time=60)
+def test_edit_image_3_inpainting_insert_mask_mode() -> None:
+    response = (
+        edit_image_3_inpainting_insert_mask_mode.edit_image_3_inpainting_insert_mask_mode(
+            _INPUT_FILE,
+            _MASK_MODE,
+            _OUTPUT_FILE,
+            _PROMPT,
+        )
+    )
+
+    assert len(response[0]._image_bytes) > 1000

generative_ai/image_generation/imagen3_editing/edit_image_3_inpainting_insert_mask_test.py

@@ -0,0 +1,39 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import os
+
+import backoff
+
+import edit_image_3_inpainting_insert_mask
+
+from google.api_core.exceptions import ResourceExhausted
+
+
+_RESOURCES = os.path.join(os.path.dirname(__file__), "../test_resources")
+_INPUT_FILE = os.path.join(_RESOURCES, "woman.png")
+_MASK_FILE = os.path.join(_RESOURCES, "woman_inpainting_insert_mask.png")
+_OUTPUT_FILE = os.path.join(_RESOURCES, "woman_with_hat.png")
+_PROMPT = "hat"
+
+
+@backoff.on_exception(backoff.expo, ResourceExhausted, max_time=60)
+def test_edit_image_3_inpainting_insert_mask() -> None:
+    response = edit_image_3_inpainting_insert_mask.edit_image_3_inpainting_insert_mask(
+        _INPUT_FILE,
+        _MASK_FILE,
+        _OUTPUT_FILE,
+        _PROMPT,
+    )
+
+    assert len(response[0]._image_bytes) > 1000

generative_ai/image_generation/imagen3_editing/edit_image_3_inpainting_remove_mask.py

@@ -0,0 +1,81 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Google Cloud Vertex AI sample for editing an image using a mask file.
+    Inpainting can remove an object from the masked area.
+"""
+import os
+
+from vertexai.preview import vision_models
+
+PROJECT_ID = os.getenv("GOOGLE_CLOUD_PROJECT")
+
+
+def edit_image_3_inpainting_remove_mask(
+    input_file: str,
+    mask_file: str,
+    output_file: str,
+    prompt: str,
+) -> vision_models.ImageGenerationResponse:
+    # [START generativeaionvertexai_imagen_edit_image_3_inpainting_remove_mask]
+
+    import vertexai
+    from vertexai.preview.vision_models import Image, ImageGenerationModel, MaskReferenceImage, RawReferenceImage
+
+    # TODO(developer): Update and un-comment below lines
+    # PROJECT_ID = "your-project-id"
+    # input_file = "input-image.png"
+    # mask_file = "mask-image.png"
+    # output_file = "output-image.png"
+    # prompt = "" # The text prompt describing the entire image.
+
+    vertexai.init(project=PROJECT_ID, location="us-central1")
+
+    model = ImageGenerationModel.from_pretrained("imagen-3.0-capability-001")
+    base_img = Image.load_from_file(location=input_file)
+    mask_img = Image.load_from_file(location=mask_file)
+    base_ref_image = RawReferenceImage(image=base_img, reference_id=0)
+    mask_ref_image = MaskReferenceImage(
+        reference_id=1, image=mask_img, mask_mode="user_provided"
+    )
+
+    images = model.edit_image(
+        reference_images=[base_ref_image, mask_ref_image],
+        prompt=prompt,
+        edit_mode="inpainting-remove",
+        # Optional parameters
+        # negative_prompt="", # Describes the object being removed (i.e., "person")
+    )
+
+    images[0].save(location=output_file, include_generation_parameters=False)
+
+    # Optional. View the edited image in a notebook.
+    # images[0].show()
+
+    print(f"Created output image using {len(images[0]._image_bytes)} bytes")
+    # Example response:
+    # Created output image using 12345678 bytes
+
+    # [END generativeaionvertexai_imagen_edit_image_3_inpainting_remove_mask]
+
+    return images
+
+
+if __name__ == "__main__":
+    edit_image_3_inpainting_remove_mask(
+        input_file="../test_resources/volleyball_game.png",
+        mask_file="../test_resources/volleyball_game_inpainting_remove_mask.png",
+        output_file="../test_resources/volleyball_game_single_blue_player.png",
+        prompt="volleyball game",
+    )