Break-A-Scene: Extracting Multiple Concepts from a Single Image

Omri Avrahami1,2, Kfir Aberman1, Ohad Fried4, Daniel Cohen-Or1,3, Dani Lischinski1,2
1Google Research, 2The Hebrew University of Jerusalem, 3Tel Aviv University, 4Reichman University

SIGGRAPH Asia 2023

PDF arXiv Video Code
Break-A-Scene teaser.

Break-A-Scene: Given a single image with multiple concepts, annotated by loose segmentation masks, our method can learn a distinct token for each concept, and use natural language guidance to re-synthesize the individual concepts or combinations of them in various contexts.

Abstract

Text-to-image model personalization aims to introduce a user-provided concept to the model, allowing its synthesis in diverse contexts. However, current methods primarily focus on the case of learning a single concept from multiple images with variations in backgrounds and poses, and struggle when adapted to a different scenario. In this work, we introduce the task of textual scene decomposition: given a single image of a scene that may contain several concepts, we aim to extract a distinct text token for each concept, enabling fine-grained control over the generated scenes. To this end, we propose augmenting the input image with masks that indicate the presence of target concepts. These masks can be provided by the user or generated automatically by a pre-trained segmentation model. We then present a novel two-phase customization process that optimizes a set of dedicated textual embeddings (handles), as well as the model weights, striking a delicate balance between accurately capturing the concepts and avoiding overfitting. We employ a masked diffusion loss to enable handles to generate their assigned concepts, complemented by a novel loss on cross-attention maps to prevent entanglement. We also introduce union-sampling, a training strategy aimed to improve the ability of combining multiple concepts in generated images. We use several automatic metrics to quantitatively compare our method against several baselines, and further affirm the results using a user study. Finally, we showcase several applications of our method.

Video

Method

Humans have a natural ability to decompose complex scenes into their constituent parts and envision them in diverse contexts. For instance, given a photo of a ceramic artwork depicting a creature seated on a bowl, one can effortlessly imagine the same creature in a variety of different poses and locations, or envision the same bowl in a new setting. However, today's generative models struggle when confronted with this type of task.


In this work, we propose a novel customization pipeline that effectively balances the preservation of learned concept identity with the avoidance of overfitting. Our pipeline, depicted above, consists of two phases. In the first phase, we designate a set of dedicated text tokens (handles), freeze the model weights, and optimize the handles to reconstruct the input image. In the second phase, we switch to fine-tuning the model weights, while continuing to optimize the handles.

We also recognize that in order to generate images exhibiting combinations of concepts, the customization process cannot be carried out separately for each concept. This observation leads us to introduce union-sampling, a training strategy that addresses this requirement and enhances the generation of concept combinations.

A crucial focus of our approach is on disentangled concept extraction, i.e., ensuring that each handle is associated with only a single target concept. To achieve this, we employ a masked version of the standard diffusion loss, which guarantees that each custom handle can generate its designated concept; however, this loss does not penalize the model for associating a handle with multiple concepts. Our main insight is that we can penalize such entanglement by additionally imposing a loss on the cross-attention maps, known to correlate with the scene layout. This additional loss ensures that each handle attends only to the areas covered by its target concept.

Results

Scene Breaking Examples

Here are some results of scene breaking and re-synthesis in different contexts and combinations:


Entangled Scene Decomposition

Our method is also capable of breaking entangled scenes:






Local Image editing

Our incorporate our method with Blended Latent Diffusion to achieve local image editing by example:








Background Extraction

Our method is also capable of extracting the background of a scene for further edits:


Image variations

Given a single image containing multiple concepts of interest, once these are extracted using our method, they can be used to generate multiple variations of the image. The arrangement of the objects in the scene, as well as the background, are different in each generation.








BibTeX

If you find this project useful for your research, please cite the following:

@inproceedings{avrahami2023bas,
  author = {Avrahami, Omri and Aberman, Kfir and Fried, Ohad and Cohen-Or, Daniel and Lischinski, Dani},
  title = {Break-A-Scene: Extracting Multiple Concepts from a Single Image},
  year = {2023},
  isbn = {9798400703157},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  url = {https://doi.org/10.1145/3610548.3618154},
  doi = {10.1145/3610548.3618154},        
  booktitle = {SIGGRAPH Asia 2023 Conference Papers},
  articleno = {96},
  numpages = {12},
  keywords = {textual inversion, multiple concept extraction, personalization},
  location = {, Sydney, NSW, Australia, },
  series = {SA '23}
}