Novel view synthesis is a long-standing problem. In this work, we consider a variant of the problem where we are given only a few context views sparsely covering a scene or an object. The goal is to predict novel viewpoints in the scene, which requires learning priors. The current state of the art is based on Neural Radiance Field (NeRF), and while achieving impressive results, the methods suffer from long training times as they require evaluating millions of 3D point samples via a neural network for each image. We propose a 2D-only method that maps multiple context views and a query pose to a new image in a single pass of a neural network. Our model uses a two-stage architecture consisting of a codebook and a transformer model. The codebook is used to embed individual images into a smaller latent space, and the transformer solves the view synthesis task in this more compact space. To train our model efficiently, we introduce a novel branching attention mechanism that allows us to use the same model not only for neural rendering but also for camera pose estimation. Experimental results on real-world scenes show that our approach is competitive compared to NeRF-based methods while not reasoning explicitly in 3D, and it is faster to train.
相關內容
Recently, Neural Radiance Fields (NeRF) has shown promising performances on reconstructing 3D scenes and synthesizing novel views from a sparse set of 2D images. Albeit effective, the performance of NeRF is highly influenced by the quality of training samples. With limited posed images from the scene, NeRF fails to generalize well to novel views and may collapse to trivial solutions in unobserved regions. This makes NeRF impractical under resource-constrained scenarios. In this paper, we present a novel learning framework, ActiveNeRF, aiming to model a 3D scene with a constrained input budget. Specifically, we first incorporate uncertainty estimation into a NeRF model, which ensures robustness under few observations and provides an interpretation of how NeRF understands the scene. On this basis, we propose to supplement the existing training set with newly captured samples based on an active learning scheme. By evaluating the reduction of uncertainty given new inputs, we select the samples that bring the most information gain. In this way, the quality of novel view synthesis can be improved with minimal additional resources. Extensive experiments validate the performance of our model on both realistic and synthetic scenes, especially with scarcer training data. Code will be released at \url{//github.com/LeapLabTHU/ActiveNeRF}.
We have recently seen tremendous progress in the neural advances for photo-real human modeling and rendering. However, it's still challenging to integrate them into an existing mesh-based pipeline for downstream applications. In this paper, we present a comprehensive neural approach for high-quality reconstruction, compression, and rendering of human performances from dense multi-view videos. Our core intuition is to bridge the traditional animated mesh workflow with a new class of highly efficient neural techniques. We first introduce a neural surface reconstructor for high-quality surface generation in minutes. It marries the implicit volumetric rendering of the truncated signed distance field (TSDF) with multi-resolution hash encoding. We further propose a hybrid neural tracker to generate animated meshes, which combines explicit non-rigid tracking with implicit dynamic deformation in a self-supervised framework. The former provides the coarse warping back into the canonical space, while the latter implicit one further predicts the displacements using the 4D hash encoding as in our reconstructor. Then, we discuss the rendering schemes using the obtained animated meshes, ranging from dynamic texturing to lumigraph rendering under various bandwidth settings. To strike an intricate balance between quality and bandwidth, we propose a hierarchical solution by first rendering 6 virtual views covering the performer and then conducting occlusion-aware neural texture blending. We demonstrate the efficacy of our approach in a variety of mesh-based applications and photo-realistic free-view experiences on various platforms, i.e., inserting virtual human performances into real environments through mobile AR or immersively watching talent shows with VR headsets.
Can language models (LM) ground question-answering (QA) tasks in the knowledge base via inherent relational reasoning ability? While previous models that use only LMs have seen some success on many QA tasks, more recent methods include knowledge graphs (KG) to complement LMs with their more logic-driven implicit knowledge. However, effectively extracting information from structured data, like KGs, empowers LMs to remain an open question, and current models rely on graph techniques to extract knowledge. In this paper, we propose to solely leverage the LMs to combine the language and knowledge for knowledge based question-answering with flexibility, breadth of coverage and structured reasoning. Specifically, we devise a knowledge construction method that retrieves the relevant context with a dynamic hop, which expresses more comprehensivenes than traditional GNN-based techniques. And we devise a deep fusion mechanism to further bridge the information exchanging bottleneck between the language and the knowledge. Extensive experiments show that our model consistently demonstrates its state-of-the-art performance over CommensenseQA benchmark, showcasing the possibility to leverage LMs solely to robustly ground QA into the knowledge base.
Point-interactive image colorization aims to colorize grayscale images when a user provides the colors for specific locations. It is essential for point-interactive colorization methods to appropriately propagate user-provided colors (i.e., user hints) in the entire image to obtain a reasonably colorized image with minimal user effort. However, existing approaches often produce partially colorized results due to the inefficient design of stacking convolutional layers to propagate hints to distant relevant regions. To address this problem, we present iColoriT, a novel point-interactive colorization Vision Transformer capable of propagating user hints to relevant regions, leveraging the global receptive field of Transformers. The self-attention mechanism of Transformers enables iColoriT to selectively colorize relevant regions with only a few local hints. Our approach colorizes images in real-time by utilizing pixel shuffling, an efficient upsampling technique that replaces the decoder architecture. Also, in order to mitigate the artifacts caused by pixel shuffling with large upsampling ratios, we present the local stabilizing layer. Extensive quantitative and qualitative results demonstrate that our approach highly outperforms existing methods for point-interactive colorization, producing accurately colorized images with a user's minimal effort. Official codes are available at //pmh9960.github.io/research/iColoriT
Creating novel views from a single image has achieved tremendous strides with advanced autoregressive models. Although recent methods generate high-quality novel views, synthesizing with only one explicit or implicit 3D geometry has a trade-off between two objectives that we call the ``seesaw'' problem: 1) preserving reprojected contents and 2) completing realistic out-of-view regions. Also, autoregressive models require a considerable computational cost. In this paper, we propose a single-image view synthesis framework for mitigating the seesaw problem. The proposed model is an efficient non-autoregressive model with implicit and explicit renderers. Motivated by characteristics that explicit methods well preserve reprojected pixels and implicit methods complete realistic out-of-view region, we introduce a loss function to complement two renderers. Our loss function promotes that explicit features improve the reprojected area of implicit features and implicit features improve the out-of-view area of explicit features. With the proposed architecture and loss function, we can alleviate the seesaw problem, outperforming autoregressive-based state-of-the-art methods and generating an image $\approx$100 times faster. We validate the efficiency and effectiveness of our method with experiments on RealEstate10K and ACID datasets.
Vision Transformers (ViTs) have proven to be effective, in solving 2D image understanding tasks by training over large-scale image datasets; and meanwhile as a somehow separate track, in modeling the 3D visual world too such as voxels or point clouds. However, with the growing hope that transformers can become the "universal" modeling tool for heterogeneous data, ViTs for 2D and 3D tasks have so far adopted vastly different architecture designs that are hardly transferable. That invites an (over-)ambitious question: can we close the gap between the 2D and 3D ViT architectures? As a piloting study, this paper demonstrates the appealing promise to understand the 3D visual world, using a standard 2D ViT architecture, with only minimal customization at the input and output levels without redesigning the pipeline. To build a 3D ViT from its 2D sibling, we "inflate" the patch embedding and token sequence, accompanied with new positional encoding mechanisms designed to match the 3D data geometry. The resultant "minimalist" 3D ViT, named Simple3D-Former, performs surprisingly robustly on popular 3D tasks such as object classification, point cloud segmentation and indoor scene detection, compared to highly customized 3D-specific designs. It can hence act as a strong baseline for new 3D ViTs. Moreover, we note that pursing a unified 2D-3D ViT design has practical relevance besides just scientific curiosity. Specifically, we demonstrate that Simple3D-Former naturally enables to exploit the wealth of pre-trained weights from large-scale realistic 2D images (e.g., ImageNet), which can be plugged in to enhancing the 3D task performance "for free".
In this paper, we present a simple seq2seq formulation for view synthesis where we take a set of ray points as input and output colors corresponding to the rays. Directly applying a standard transformer on this seq2seq formulation has two limitations. First, the standard attention cannot successfully fit the volumetric rendering procedure, and therefore high-frequency components are missing in the synthesized views. Second, applying global attention to all rays and pixels is extremely inefficient. Inspired by the neural radiance field (NeRF), we propose the NeRF attention (NeRFA) to address the above problems. On the one hand, NeRFA considers the volumetric rendering equation as a soft feature modulation procedure. In this way, the feature modulation enhances the transformers with the NeRF-like inductive bias. On the other hand, NeRFA performs multi-stage attention to reduce the computational overhead. Furthermore, the NeRFA model adopts the ray and pixel transformers to learn the interactions between rays and pixels. NeRFA demonstrates superior performance over NeRF and NerFormer on four datasets: DeepVoxels, Blender, LLFF, and CO3D. Besides, NeRFA establishes a new state-of-the-art under two settings: the single-scene view synthesis and the category-centric novel view synthesis.
Transformer, an attention-based encoder-decoder architecture, has revolutionized the field of natural language processing. Inspired by this significant achievement, some pioneering works have recently been done on adapting Transformerliked architectures to Computer Vision (CV) fields, which have demonstrated their effectiveness on various CV tasks. Relying on competitive modeling capability, visual Transformers have achieved impressive performance on multiple benchmarks such as ImageNet, COCO, and ADE20k as compared with modern Convolution Neural Networks (CNN). In this paper, we have provided a comprehensive review of over one hundred different visual Transformers for three fundamental CV tasks (classification, detection, and segmentation), where a taxonomy is proposed to organize these methods according to their motivations, structures, and usage scenarios. Because of the differences in training settings and oriented tasks, we have also evaluated these methods on different configurations for easy and intuitive comparison instead of only various benchmarks. Furthermore, we have revealed a series of essential but unexploited aspects that may empower Transformer to stand out from numerous architectures, e.g., slack high-level semantic embeddings to bridge the gap between visual and sequential Transformers. Finally, three promising future research directions are suggested for further investment.
Generating texts which express complex ideas spanning multiple sentences requires a structured representation of their content (document plan), but these representations are prohibitively expensive to manually produce. In this work, we address the problem of generating coherent multi-sentence texts from the output of an information extraction system, and in particular a knowledge graph. Graphical knowledge representations are ubiquitous in computing, but pose a significant challenge for text generation techniques due to their non-hierarchical nature, collapsing of long-distance dependencies, and structural variety. We introduce a novel graph transforming encoder which can leverage the relational structure of such knowledge graphs without imposing linearization or hierarchical constraints. Incorporated into an encoder-decoder setup, we provide an end-to-end trainable system for graph-to-text generation that we apply to the domain of scientific text. Automatic and human evaluations show that our technique produces more informative texts which exhibit better document structure than competitive encoder-decoder methods.
In this paper, we propose the joint learning attention and recurrent neural network (RNN) models for multi-label classification. While approaches based on the use of either model exist (e.g., for the task of image captioning), training such existing network architectures typically require pre-defined label sequences. For multi-label classification, it would be desirable to have a robust inference process, so that the prediction error would not propagate and thus affect the performance. Our proposed model uniquely integrates attention and Long Short Term Memory (LSTM) models, which not only addresses the above problem but also allows one to identify visual objects of interests with varying sizes without the prior knowledge of particular label ordering. More importantly, label co-occurrence information can be jointly exploited by our LSTM model. Finally, by advancing the technique of beam search, prediction of multiple labels can be efficiently achieved by our proposed network model.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191