We tackle the problem of sparse novel view synthesis (NVS) using video diffusion models; given $K$ ($\approx 5$) multi-view images of a scene and their camera poses, we predict the view from a target camera pose. Many prior approaches leverage generative image priors encoded via diffusion models. However, models trained on single images lack multi-view knowledge. We instead argue that video models already contain implicit multi-view knowledge and so should be easier to adapt for NVS. Our key insight is to formulate sparse NVS as a low frame-rate video completion task. However, one challenge is that sparse NVS is defined over an unordered set of inputs, often too sparse to admit a meaningful order, so the models should be $\textit{invariant}$ to permutations of that input set. To this end, we present FrameCrafter, which adapts video models (naturally trained with coherent frame orderings) to permutation-invariant NVS through several architectural modifications, including per-frame latent encodings and removal of temporal positional embeddings. Our results suggest that video models can be easily trained to "forget" about time with minimal supervision, producing competitive performance on sparse-view NVS benchmarks. Project page: https://frame-crafter.github.io/
ST Engineering - NTU Corporate Lab（ST工程-南洋理工大学企业实验室）；Nanyang Technological University（南洋理工大学）；Department of Computer Science, City University of Hong Kong（香港城市大学计算机科学系）；City University of Hong Kong（香港城市大学）
The light field faithfully records the spatial and angular configurations of the scene, which facilitates a wide range of imaging possibilities. In this work, we propose an LF synthesis algorithm which renders high quality novel LF views far outside the range of angular baselines of the given references. A stratified synthesis strategy is adopted which parses the scene content based on stratified disparity layers and across a varying range of spatial granularities. Such a stratified methodology proves to help preserve scene structures over large perspective shifts, and it provides informative …
Generative Adversarial Networks (GANs) have proven successful for unsupervised image generation. Several works have extended GANs to image inpainting by conditioning the generation with parts of the image to be reconstructed. Despite their success, these methods have limitations in settings where only a small subset of the image pixels is known beforehand. In this paper we investigate the effectiveness of conditioning GANs when very few pixel values are provided. We propose a modelling framework which results in adding an explicit cost term to the GAN objective function to enforce pixel-wise c…
GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space
Shanghai Jiao Tong University（上海交通大学）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；The Chinese University of Hong Kong（香港中文大学）；The University of Hong Kong（香港大学）；Fudan University（复旦大学）；Zhejiang University（浙江大学）
The goal of this thesis is to present my research contributions towards solving various visual synthesis and generation tasks, comprising image translation, image completion, and completed scene decomposition. This thesis consists of five pieces of work, each of which presents a new learning-based approach for synthesizing images with plausible content as well as visually realistic appearance. Each work demonstrates the superiority of the proposed approach on image synthesis, with some further contributing to other tasks, such as depth estimation.
University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）；Adobe Research（Adobe研究院）
We present a transformation-grounded image generation network for novel 3D view synthesis from a single image. Instead of taking a 'blank slate' approach, we first explicitly infer the parts of the geometry visible both in the input and novel views and then re-cast the remaining synthesis problem as image completion. Specifically, we both predict a flow to move the pixels from the input to the novel view along with a novel visibility map that helps deal with occulsion/disocculsion. Next, conditioned on those intermediate results, we hallucinate (infer) parts of the object invisible in the inpu…
We present 3DiM, a diffusion model for 3D novel view synthesis, which is able to translate a single input view into consistent and sharp completions across many views. The core component of 3DiM is a pose-conditional image-to-image diffusion model, which takes a source view and its pose as inputs, and generates a novel view for a target pose as output. 3DiM can generate multiple views that are 3D consistent using a novel technique called stochastic conditioning. The output views are generated autoregressively, and during the generation of each novel view, one selects a random conditioning view…
Tsinghua University（清华大学）；Tsinghua Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；University College London（伦敦大学学院）
Image generation has received increasing attention because of its wide application in security and entertainment. Sketch-based face generation brings more fun and better quality of image generation due to supervised interaction. However, when a sketch poorly aligned with the true face is given as input, existing supervised image-to-image translation methods often cannot generate acceptable photo-realistic face images. To address this problem, in this paper we propose Cali-Sketch, a human-like-sketch to photo-realistic-image generation method. Cali-Sketch explicitly models stroke calibration an…
Generative modeling of human motion has broad applications in computer animation, virtual reality, and robotics. Conventional approaches develop separate models for different motion synthesis tasks, and typically use a model of a small size to avoid overfitting the scarce data available in each setting. It remains an open question whether developing a single unified model is feasible, which may 1) benefit the acquirement of novel skills by combining skills learned from multiple tasks, and 2) help in increasing the model capacity without overfitting by combining multiple data sources. Unificati…
Existing face completion solutions are primarily driven by end-to-end models that directly generate 2D completions of 2D masked faces. By having to implicitly account for geometric and photometric variations in facial shape and appearance, such approaches result in unrealistic completions, especially under large variations in pose, shape, illumination and mask sizes. To alleviate these limitations, we introduce 3DFaceFill, an analysis-by-synthesis approach for face completion that explicitly considers the image formation process. It comprises three components, (1) an encoder that disentangles …
Stanford University（斯坦福大学）；Technical University of Munich（慕尼黑工业大学）
We introduce a data-driven approach to complete partial 3D shapes through a combination of volumetric deep neural networks and 3D shape synthesis. From a partially-scanned input shape, our method first infers a low-resolution -- but complete -- output. To this end, we introduce a 3D-Encoder-Predictor Network (3D-EPN) which is composed of 3D convolutional layers. The network is trained to predict and fill in missing data, and operates on an implicit surface representation that encodes both known and unknown space. This allows us to predict global structure in unknown areas at high accuracy. We …
View synthesis is a process for generating novel views from a scene which has been recorded with a 3-D camera setup. It has important applications in 3-D post-production and 2-D to 3-D conversion. However, a central problem in the generation of novel views lies in the handling of disocclusions. Background content, which was occluded in the original view, may become unveiled in the synthesized view. This leads to missing information in the generated view which has to be filled in a visually plausible manner. We present an inpainting algorithm for disocclusion filling in synthesized views based …
Technical University of Munich（慕尼黑工业大学）；Meta Reality Labs Zurich（Meta苏黎世现实实验室）
We introduce DiffRF, a novel approach for 3D radiance field synthesis based on denoising diffusion probabilistic models. While existing diffusion-based methods operate on images, latent codes, or point cloud data, we are the first to directly generate volumetric radiance fields. To this end, we propose a 3D denoising model which directly operates on an explicit voxel grid representation. However, as radiance fields generated from a set of posed images can be ambiguous and contain artifacts, obtaining ground truth radiance field samples is non-trivial. We address this challenge by pairing the d…
SynergyAmodal: Deocclude Anything with Text Control
Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, Xiamen University, China（多媒体可信感知与高效计算重点实验室、教育部、厦门大学）
City University of Hong Kong（香港城市大学）；The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））；Zhejiang University（浙江大学）
Recent deep generative models allow real-time generation of hair images from sketch inputs. Existing solutions often require a user-provided binary mask to specify a target hair shape. This not only costs users extra labor but also fails to capture complicated hair boundaries. Those solutions usually encode hair structures via orientation maps, which, however, are not very effective to encode complex structures. We observe that colored hair sketches already implicitly define target hair shapes as well as hair appearance and are more flexible to depict hair structures than orientation maps. Bas…
CVSformer: Cross-View Synthesis Transformer for Semantic Scene Completion
Tianjin University（天津大学）；Shenzhen University（深圳大学）；A*STAR（新加坡科技研究局）；The Hong Kong Polytechnic University（香港理工大学）；South China University of Technology（华南理工大学）；Alibaba Damo Academy（阿里巴巴达摩院）
Zhejiang University（浙江大学）；The University of Tokyo（东京大学）；Microsoft（微软）；University of Amsterdam（阿姆斯特丹大学）
We introduce a scalable framework for novel view synthesis from RGB-D images with largely incomplete scene coverage. While generative neural approaches have demonstrated spectacular results on 2D images, they have not yet achieved similar photorealistic results in combination with scene completion where a spatial 3D scene understanding is essential. To this end, we propose a generative pipeline performing on a sparse grid-based neural scene representation to complete unobserved scene parts via a learned distribution of scenes in a 2.5D-3D-2.5D manner. We process encoded image features in 3D sp…
Inpainting high-resolution images with large holes challenges existing deep learning based image inpainting methods. We present a novel framework -- PyramidFill for high-resolution image inpainting task, which explicitly disentangles content completion and texture synthesis. PyramidFill attempts to complete the content of unknown regions in a lower-resolution image, and synthesis the textures of unknown regions in a higher-resolution image, progressively. Thus, our model consists of a pyramid of fully convolutional GANs, wherein the content GAN is responsible for completing contents in the low…
Swapping text in scene images while preserving original fonts, colors, sizes and background textures is a challenging task due to the complex interplay between different factors. In this work, we present SwapText, a three-stage framework to transfer texts across scene images. First, a novel text swapping network is proposed to replace text labels only in the foreground image. Second, a background completion network is learned to reconstruct background images. Finally, the generated foreground image and background image are used to generate the word image by the fusion network. Using the propos…
Probabilistic diffusion models have achieved state-of-the-art results for image synthesis, inpainting, and text-to-image tasks. However, they are still in the early stages of generating complex 3D shapes. This work proposes Diffusion-SDF, a generative model for shape completion, single-view reconstruction, and reconstruction of real-scanned point clouds. We use neural signed distance functions (SDFs) as our 3D representation to parameterize the geometry of various signals (e.g., point clouds, 2D images) through neural networks. Neural SDFs are implicit functions and diffusing them amounts to l…
Pattern-Aware Diffusion Synthesis of fMRI/dMRI with Tissue and Microstructural Refinement
Department of Computer Science and Technology, Harbin Institute of Technology（计算机科学与技术系，哈尔滨工业大学）；School of Intelligence Science and Engineering, College of Artificial Intelligence, Harbin Institute of Technology（智能科学与工程学院，人工智能学院，哈尔滨工业大学）；School of Biomedical Engineering, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, Guangdong Key Laboratory for Biomedical, Measurements and Ultrasound Imaging, Shenzhen University Medical School, Shenzhen University, Shenzhen, China（生物医学工程学院，医学超声关键技术工程实验室，广东生物医学、测量与超声成像重点实验室，深圳大学医学院，深圳大学，深圳，中国）；Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences（深圳先进技术研究院，中国科学院）；Department of Radiology, The People’s Hospital of Guangxi Zhuang Autonomous Region, Guangxi Academy of Medical Sciences（放射科，广西壮族自治区人民医院，广西医学院）
Adobe Research India（Adobe印度研究院）；Arizona State University（亚利桑那州立大学）；Carnegie Mellon University（卡内基梅隆大学）
Recently, there has been substantial progress in image synthesis from semantic labelmaps. However, methods used for this task assume the availability of complete and unambiguous labelmaps, with instance boundaries of objects, and class labels for each pixel. This reliance on heavily annotated inputs restricts the application of image synthesis techniques to real-world applications, especially under uncertainty due to weather, occlusion, or noise. On the other hand, algorithms that can synthesize images from sparse labelmaps or sketches are highly desirable as tools that can guide content creat…
National Laboratory of Pattern Recognition, CASIA（中国科学院自动化研究所模式识别国家重点实验室）；Center for Research on Intelligent Perception and Computing, CASIA（中国科学院自动化研究所智能感知与计算研究中心）；Center for Excellence in Brain Science and Intelligence Technology, CAS（中国科学院脑科学与智能技术卓越创新中心）；University of Chinese Academy of Sciences（中国科学院大学）
Near infrared-visible (NIR-VIS) heterogeneous face recognition refers to the process of matching NIR to VIS face images. Current heterogeneous methods try to extend VIS face recognition methods to the NIR spectrum by synthesizing VIS images from NIR images. However, due to self-occlusion and sensing gap, NIR face images lose some visible lighting contents so that they are always incomplete compared to VIS face images. This paper models high resolution heterogeneous face synthesis as a complementary combination of two components, a texture inpainting component and pose correction component. The…
University of Kentucky（肯塔基大学）；University of Southern California（南加州大学）；Harbin Institute of Technology（哈尔滨工业大学）；Hangzhou Normal University（杭州师范大学）
We present a novel deep learning approach to synthesize complete face images in the presence of large ocular region occlusions. This is motivated by recent surge of VR/AR displays that hinder face-to-face communications. Different from the state-of-the-art face inpainting methods that have no control over the synthesized content and can only handle frontal face pose, our approach can faithfully recover the missing content under various head poses while preserving the identity. At the core of our method is a novel generative network with dedicated constraints to regularize the synthesis process…
ShanghaiTech University（上海科技大学）；Shanghai United Imaging Intelligence Co., Ltd.（上海联影智能科技有限公司）；Northwestern Polytechnical University（西北工业大学）
Multi-modal medical image completion has been extensively applied to alleviate the missing modality issue in a wealth of multi-modal diagnostic tasks. However, for most existing synthesis methods, their inferences of missing modalities can collapse into a deterministic mapping from the available ones, ignoring the uncertainties inherent in the cross-modal relationships. Here, we propose the Unified Multi-Modal Conditional Score-based Generative Model (UMM-CSGM) to take advantage of Score-based Generative Model (SGM) in modeling and stochastically sampling a target probability distribution, and…
We present a new, fast and flexible pipeline for indoor scene synthesis that is based on deep convolutional generative models. Our method operates on a top-down image-based representation, and inserts objects iteratively into the scene by predicting their category, location, orientation and size with separate neural network modules. Our pipeline naturally supports automatic completion of partial scenes, as well as synthesis of complete scenes. Our method is significantly faster than the previous image-based method and generates result that outperforms it and other state-of-the-art deep generat…
University of Amsterdam（阿姆斯特丹大学）；Google Research（谷歌研究院）
In this paper the argument is made that for true novel view synthesis of objects, where the object can be synthesized from any viewpoint, an explicit 3D shape representation isdesired. Our method estimates point clouds to capture the geometry of the object, which can be freely rotated into the desired view and then projected into a new image. This image, however, is sparse by nature and hence this coarse view is used as the input of an image completion network to obtain the dense target view. The point cloud is obtained using the predicted pixel-wise depth map, estimated from a single RGB inpu…
InFusion: Inpainting 3D Gaussians via Learning Depth Completion from Diffusion Prior
University of Science and Technology of China（中国科学技术大学）；Alibaba Group（阿里巴巴集团）；The Hong Kong University of Science and Technology（香港科技大学）；Ant Group（蚂蚁集团）
University of North Georgia（北乔治亚大学）；Lawrence Technological University（劳伦斯理工大学）；University Massachusetts Dartmouth（麻省大学达特茅斯分校）；South Dakota State University（南达科他州立大学）；University of Kentucky（肯塔基大学）
With the explosive growth of web-based cameras and mobile devices, billions of photographs are uploaded to the internet. We can trivially collect a huge number of photo streams for various goals, such as 3D scene reconstruction and other big data applications. However, this is not an easy task due to the fact the retrieved photos are neither aligned nor calibrated. Furthermore, with the occlusion of unexpected foreground objects like people, vehicles, it is even more challenging to find feature correspondences and reconstruct realistic scenes. In this paper, we propose a structure based image …
