Abstract:As a common image editing operation, image composition aims to combine the foreground from one image and another background image, resulting in a composite image. However, there are many issues that could make the composite images unrealistic. These issues can be summarized as the inconsistency between foreground and background, which includes appearance inconsistency (e.g., incompatible illumination), geometry inconsistency (e.g., unreasonable size), and semantic inconsistency (e.g., mismatched semantic context). Image composition task could be decomposed into multiple sub-tasks, in which each sub-task targets at one or more issues. Specifically, object placement aims to find reasonable scale, location, and shape for the foreground. Image blending aims to address the unnatural boundary between foreground and background. Image harmonization aims to adjust the illumination statistics of foreground. Shadow generation aims to generate plausible shadow for the foreground. These sub-tasks can be executed sequentially or parallelly to acquire realistic composite images. To the best of our knowledge, there is no previous survey on image composition. In this paper, we conduct comprehensive survey over the sub-tasks and combinatorial task of image composition. For each one, we summarize the existing methods, available datasets, and common evaluation metrics. Datasets and codes for image composition are summarized at <a class="link-external link-https" href="https://github.com/bcmi/Awesome-Image-Composition" rel="external noopener nofollow">this https URL</a>. We have also contributed the first image composition toolbox: libcom <a class="link-external link-https" href="https://github.com/bcmi/libcom" rel="external noopener nofollow">this https URL</a>, which assembles 10+ image composition related functions (e.g., image blending, image harmonization, object placement, shadow generation, generative composition). The ultimate goal of this toolbox is solving all the problems related to image composition with simple `import libcom'.

A review of recent image synthesis techniques

Making Images Real Again: A Comprehensive Survey on Deep Image Composition

Deep image synthesis from intuitive user input: A review and perspectives

A Deep Learning Network Based End-to-end Image Composition.

Image synthesis: a review of methods, datasets, evaluation metrics, and future outlook

Novel 3D-Aware Composition Images Synthesis for Object Display with Diffusion Model.

Recent Progress of Face Image Synthesis

Scene Text Synthesis for Efficient and Effective Deep Network Training

Image Composition Method Based on a Spatial Position Analysis Network

Creative and High-Quality Image Composition Based on a New Criterion

Image Synthesis under Limited Data: A Survey and Taxonomy

Hierarchy Composition GAN for High-fidelity Image Synthesis

A Survey on Deep Generative 3D-aware Image Synthesis

A State-of-the-Art Review on Image Synthesis With Generative Adversarial Networks

View-Aware Image Object Compositing and Synthesis from Multiple Sources

Text-to-image synthesis: Starting composite from the foreground content

Composition-Aided Face Photo-Sketch Synthesis.

A Review on Medical Imaging Synthesis Using Deep Learning and Its Clinical Applications

Human Image Generation: A Comprehensive Survey

A survey of generative adversarial networks and their application in text-to-image synthesis

FreeCompose: Generic Zero-Shot Image Composition with Diffusion Prior