Separate-and-Enhance: Compositional Finetuning for Text2Image Diffusion Models

Despite recent significant strides achieved by diffusion-based Text-to-Image (T2I) models, current systems are still less capable of ensuring decent compositional generation aligned with text prompts, particularly for the multi-object generation. This work illuminates the fundamental reasons for such misalignment, pinpointing issues related to low attention activation scores and mask overlaps. While previous research efforts have individually tackled these issues, we assert that a holistic approach is paramount. Thus, we propose two novel objectives, the Separate loss and the Enhance loss, that reduce object mask overlaps and maximize attention scores, respectively. Our method diverges from conventional test-time-adaptation techniques, focusing on finetuning critical parameters, which enhances scalability and generalizability. Comprehensive evaluations demonstrate the superior performance of our model in terms of image realism, text-image alignment, and adaptability, notably outperforming prominent baselines. Ultimately, this research paves the way for T2I diffusion models with enhanced compositional capacities and broader applicability. The project webpage is available at https://zpbao.github.io/projects/SepEn/.

Consistent Multimodal Generation via A Unified GAN Framework

We investigate how to generate multimodal image outputs, such as RGB, depth, and surface normals, with a single generative model. The challenge is to produce outputs that are realistic, and also consistent with each other. Our solution builds on the StyleGAN3 architecture, with a shared backbone and modality-specific branches in the last layers of the synthesis network, and we propose per-modality fidelity discriminators and a cross-modality consistency discriminator. In experiments on the Stanford2D3D dataset, we demonstrate realistic and consistent generation of RGB, depth, and normal images. We also show a training recipe to easily extend our pretrained model on a new domain, even with a few pairwise data. We further evaluate the use of synthetically generated RGB and depth pairs for training or fine-tuning depth estimators. Code will be available at https://github.com/jessemelpolio/MultimodalGAN.

Putting People in Their Place: Affordance-Aware Human Insertion into Scenes

We study the problem of inferring scene affordances by presenting a method for realistically inserting people into scenes. Given a scene image with a marked region and an image of a person, we insert the person into the scene while respecting the scene affordances. Our model can infer the set of realistic poses given the scene context, re-pose the reference person, and harmonize the composition. We set up the task in a self-supervised fashion by learning to re-pose humans in video clips. We train a large-scale diffusion model on a dataset of 2.4M video clips that produces diverse plausible poses while respecting the scene context. Given the learned human-scene composition, our model can also hallucinate realistic people and scenes when prompted without conditioning and also enables interactive editing. A quantitative evaluation shows that our method synthesizes more realistic human appearance and more natural human-scene interactions than prior work.

Towards Enhanced Controllability of Diffusion Models

Denoising Diffusion models have shown remarkable capabilities in generating realistic, high-quality and diverse images. However, the extent of controllability during generation is underexplored. Inspired by techniques based on GAN latent space for image manipulation, we train a diffusion model conditioned on two latent codes, a spatial content mask and a flattened style embedding. We rely on the inductive bias of the progressive denoising process of diffusion models to encode pose/layout information in the spatial structure mask and semantic/style information in the style code. We propose two generic sampling techniques for improving controllability. We extend composable diffusion models to allow for some dependence between conditional inputs, to improve the quality of generations while also providing control over the amount of guidance from each latent code and their joint distribution. We also propose timestep dependent weight scheduling for content and style latents to further improve the translations. We observe better controllability compared to existing methods and show that without explicit training objectives, diffusion models can be used for effective image manipulation and image translation.

Modulating Pretrained Diffusion Models for Multimodal Image Synthesis

We present multimodal conditioning modules (MCM) for enabling conditional image synthesis using pretrained diffusion models. Previous multimodal synthesis works rely on training networks from scratch or fine-tuning pretrained networks, both of which are computationally expensive for large, state-of-the-art diffusion models. Our method uses pretrained networks but does not require any updates to the diffusion network's parameters. MCM is a small module trained to modulate the diffusion network's predictions during sampling using 2D modalities (e.g., semantic segmentation maps, sketches) that were unseen during the original training of the diffusion model. We show that MCM enables user control over the spatial layout of the image and leads to increased control over the image generation process. Training MCM is cheap as it does not require gradients from the original diffusion net, consists of only $\sim$1$\%$ of the number of parameters of the base diffusion model, and is trained using only a limited number of training examples. We evaluate our method on unconditional and text-conditional models to demonstrate the improved control over the generated images and their alignment with respect to the conditioning inputs.

Zero-shot Image-to-Image Translation

Large-scale text-to-image generative models have shown their remarkable ability to synthesize diverse and high-quality images. However, it is still challenging to directly apply these models for editing real images for two reasons. First, it is hard for users to come up with a perfect text prompt that accurately describes every visual detail in the input image. Second, while existing models can introduce desirable changes in certain regions, they often dramatically alter the input content and introduce unexpected changes in unwanted regions. In this work, we propose pix2pix-zero, an image-to-image translation method that can preserve the content of the original image without manual prompting. We first automatically discover editing directions that reflect desired edits in the text embedding space. To preserve the general content structure after editing, we further propose cross-attention guidance, which aims to retain the cross-attention maps of the input image throughout the diffusion process. In addition, our method does not need additional training for these edits and can directly use the existing pre-trained text-to-image diffusion model. We conduct extensive experiments and show that our method outperforms existing and concurrent works for both real and synthetic image editing.

UMFuse: Unified Multi View Fusion for Human Editing applications

The vision community has explored numerous pose guided human editing methods due to their extensive practical applications. Most of these methods still use an image-to-image formulation in which a single image is given as input to produce an edited image as output. However, the problem is ill-defined in cases when the target pose is significantly different from the input pose. Existing methods then resort to in-painting or style transfer to handle occlusions and preserve content. In this paper, we explore the utilization of multiple views to minimize the issue of missing information and generate an accurate representation of the underlying human model. To fuse the knowledge from multiple viewpoints, we design a selector network that takes the pose keypoints and texture from images and generates an interpretable per-pixel selection map. After that, the encodings from a separate network (trained on a single image human reposing task) are merged in the latent space. This enables us to generate accurate, precise, and visually coherent images for different editing tasks. We show the application of our network on 2 newly proposed tasks - Multi-view human reposing, and Mix-and-match human image generation. Additionally, we study the limitations of single-view editing and scenarios in which multi-view provides a much better alternative.

VGFlow: Visibility guided Flow Network for Human Reposing

The task of human reposing involves generating a realistic image of a person standing in an arbitrary conceivable pose. There are multiple difficulties in generating perceptually accurate images, and existing methods suffer from limitations in preserving texture, maintaining pattern coherence, respecting cloth boundaries, handling occlusions, manipulating skin generation, etc. These difficulties are further exacerbated by the fact that the possible space of pose orientation for humans is large and variable, the nature of clothing items is highly non-rigid, and the diversity in body shape differs largely among the population. To alleviate these difficulties and synthesize perceptually accurate images, we propose VGFlow. Our model uses a visibility-guided flow module to disentangle the flow into visible and invisible parts of the target for simultaneous texture preservation and style manipulation. Furthermore, to tackle distinct body shapes and avoid network artifacts, we also incorporate a self-supervised patch-wise "realness" loss to improve the output. VGFlow achieves state-of-the-art results as observed qualitatively and quantitatively on different image quality metrics (SSIM, LPIPS, FID).

Contrastive Learning for Diverse Disentangled Foreground Generation

We introduce a new method for diverse foreground generation with explicit control over various factors. Existing image inpainting based foreground generation methods often struggle to generate diverse results and rarely allow users to explicitly control specific factors of variation (e.g., varying the facial identity or expression for face inpainting results). We leverage contrastive learning with latent codes to generate diverse foreground results for the same masked input. Specifically, we define two sets of latent codes, where one controls a pre-defined factor (``known''), and the other controls the remaining factors (``unknown''). The sampled latent codes from the two sets jointly bi-modulate the convolution kernels to guide the generator to synthesize diverse results. Experiments demonstrate the superiority of our method over state-of-the-arts in result diversity and generation controllability.

Spatially-Adaptive Multilayer Selection for GAN Inversion and Editing

Existing GAN inversion and editing methods work well for aligned objects with a clean background, such as portraits and animal faces, but often struggle for more difficult categories with complex scene layouts and object occlusions, such as cars, animals, and outdoor images. We propose a new method to invert and edit such complex images in the latent space of GANs, such as StyleGAN2. Our key idea is to explore inversion with a collection of layers, spatially adapting the inversion process to the difficulty of the image. We learn to predict the "invertibility" of different image segments and project each segment into a latent layer. Easier regions can be inverted into an earlier layer in the generator's latent space, while more challenging regions can be inverted into a later feature space. Experiments show that our method obtains better inversion results compared to the recent approaches on complex categories, while maintaining downstream editability. Please refer to our project page at https://www.cs.cmu.edu/~SAMInversion.