We present a novel approach for few-shot NeRF estimation, aimed at avoiding local artifacts and capable of efficiently reconstructing real scenes. In contrast to previous methods that rely on pre-trained modules or various data-driven priors that only work well in specific scenarios, our method is fully generic and is based on controlling the frequency of the learned signal in the Fourier domain. We observe that in NeRF learning methods, high-frequency artifacts often show up early in the optimization process, and the network struggles to correct them due to the lack of dense supervision in few-shot cases. To counter this, we introduce an explicit curriculum training procedure, which progressively adds higher frequencies throughout optimization, thus favoring global, low-frequency signals initially, and only adding details later. We represent the radiance fields using a grid-based model and introduce an efficient approach to control the frequency band of the learned signal in the Fourier domain. Therefore our method achieves faster reconstruction and better rendering quality than purely MLP-based methods. We show that our approach is general and is capable of producing high-quality results on real scenes, at a fraction of the cost of competing methods. Our method opens the door to efficient and accurate scene acquisition in the few-shot NeRF setting.

Using parts of existing models to rebuild new models, commonly termed as example-based modeling, is a classical methodology in the realm of computer graphics. Previous works mostly focus on shape composition, making them very hard to use for realistic composition of 3D objects captured from real-world scenes. This leads to combining multiple NeRFs into a single 3D scene to achieve seamless appearance blending. However, the current SeamlessNeRF method struggles to achieve interactive editing and harmonious stitching for real-world scenes due to its gradient-based strategy and grid-based representation. To this end, we present an example-based modeling method that combines multiple Gaussian fields in a point-based representation using sample-guided synthesis. Specifically, as for composition, we create a GUI to segment and transform multiple fields in real time, easily obtaining a semantically meaningful composition of models represented by 3D Gaussian Splatting (3DGS). For texture blending, due to the discrete and irregular nature of 3DGS, straightforwardly applying gradient propagation as SeamlssNeRF is not supported. Thus, a novel sampling-based cloning method is proposed to harmonize the blending while preserving the original rich texture and content. Our workflow consists of three steps: 1) real-time segmentation and transformation of 3DGS using a well-tailored GUI, 2) KNN analysis to identify boundary points in the intersecting area between the source and target models, and 3) two-phase optimization of the target model using sampling-based cloning and gradient constraints. Extensive experimental results validate that our approach significantly outperforms previous works in realistic synthesis, demonstrating its practicality.

We propose a novel zero-shot approach for keypoint detection on 3D shapes. Point-level reasoning on visual data is challenging as it requires precise localization capability, posing problems even for powerful models like DINO or CLIP. Traditional methods for 3D keypoint detection rely heavily on annotated 3D datasets and extensive supervised training, limiting their scalability and applicability to new categories or domains. In contrast, our method utilizes the rich knowledge embedded within Multi-Modal Large Language Models (MLLMs). Specifically, we demonstrate, for the first time, that pixel-level annotations used to train recent MLLMs can be exploited for both extracting and naming salient keypoints on 3D models without any ground truth labels or supervision. Experimental evaluations demonstrate that our approach achieves competitive performance on standard benchmarks compared to supervised methods, despite not requiring any 3D keypoint annotations during training. Our results highlight the potential of integrating language models for localized 3D shape understanding. This work opens new avenues for cross-modal learning and underscores the effectiveness of MLLMs in contributing to 3D computer vision challenges.

While recent advances in Gaussian Splatting have enabled fast reconstruction of high-quality 3D scenes from images, extracting accurate surface meshes remains a challenge. Current approaches extract the surface through costly post-processing steps, resulting in the loss of fine geometric details or requiring significant time and leading to very dense meshes with millions of vertices. More fundamentally, the a posteriori conversion from a volumetric to a surface representation limits the ability of the final mesh to preserve all geometric structures captured during training. We present MILo, a novel Gaussian Splatting framework that bridges the gap between volumetric and surface representations by differentiably extracting a mesh from the 3D Gaussians. We design a fully differentiable procedure that constructs the mesh—including both vertex locations and connectivity—at every iteration directly from the parameters of the Gaussians, which are the only quantities optimized during training. Our method introduces three key technical contributions: (1) a bidirectional consistency framework ensuring both representations—Gaussians and the extracted mesh—capture the same underlying geometry during training; (2) an adaptive mesh extraction process performed at each training iteration, which uses Gaussians as differentiable pivots for Delaunay triangulation; (3) a novel method for computing signed distance values from the 3D Gaussians that enables precise surface extraction while avoiding geometric erosion. Our approach can reconstruct complete scenes, including backgrounds, with state-of-the-art quality while requiring an order of magnitude fewer mesh vertices than previous methods. Due to their light weight and empty interior, our meshes are well suited for downstream applications such as physics simulations and animation. The code for our approach and an online gallery are available at https://anttwo.github.io/milo/.

Neural Radiance Fields (NeRF) achieves unprecedented performance in synthesizing novel view synthesis, utilizing multi-view consistency. When capturing multiple inputs, image signal processing (ISP) in modern cameras will independently enhance them, including exposure adjustment, color correction, local tone mapping, etc. While these processings greatly improve image quality, they often break the multi-view consistency assumption, leading to "floaters" in the reconstructed radiance fields. To address this concern without compromising visual aesthetics, we aim to first disentangle the enhancement by ISP at the NeRF training stage and re-apply user-desired enhancements to the reconstructed radiance fields at the finishing stage. Furthermore, to make the re-applied enhancements consistent between novel views, we need to perform imaging signal processing in 3D space (i.e. "3D ISP"). For this goal, we adopt the bilateral grid, a locally-affine model, as a generalized representation of ISP processing. Specifically, we optimize per-view 3D bilateral grids with radiance fields to approximate the effects of camera pipelines for each input view. To achieve user-adjustable 3D finishing, we propose to learn a low-rank 4D bilateral grid from a given single view edit, lifting photo enhancements to the whole 3D scene. We demonstrate our approach can boost the visual quality of novel view synthesis by effectively removing floaters and performing enhancements from user retouching. The source code and our data are available at: https://bilarfpro.github.io.

Radiance fields have gradually become a main representation of media. Although its appearance editing has been studied, how to achieve view-consistent recoloring in an efficient manner is still under explored. We present RecolorNeRF, a novel user-friendly color editing approach for the neural radiance fields. Our key idea is to decompose the scene into a set of pure-colored layers, forming a palette. By this means, color manipulation can be conducted by altering the color components of the palette directly. To support efficient palette-based editing, the color of each layer needs to be as representative as possible. In the end, the problem is formulated as an optimization problem, where the layers and their blending weights are jointly optimized with the NeRF itself. Extensive experiments show that our jointly-optimized layer decomposition can be used against multiple backbones and produce photo-realistic recolored novel-view renderings. We demonstrate that RecolorNeRF outperforms baseline methods both quantitatively and qualitatively for color editing even in complex real-world scenes.

Neural Radiance Fields (NeRFs) have emerged as promising digital mediums of 3D objects and scenes, sparking a surge in research to extend the editing capabilities in this domain. The task of seamless editing and merging of multiple NeRFs, resembling the “Poisson blending” in 2D image editing, remains a critical operation that is under-explored by existing work. To fill this gap, we propose SeamlessNeRF, a novel approach for seamless appearance blending of multiple NeRFs. In specific, we aim to optimize the appearance of a target radiance field in order to harmonize its merge with a source field. We propose a well-tailored optimization procedure for blending, which is constrained by 1) pinning the radiance color in the intersecting boundary area between the source and target fields and 2) maintaining the original gradient of the target. Extensive experiments validate that our approach can effectively propagate the source appearance from the boundary area to the entire target field through the gradients. To the best of our knowledge, SeamlessNeRF is the first work that introduces gradient-guided appearance editing to radiance fields, offering solutions for seamless stitching of 3D objects represented in NeRFs. Our code and more results are available at https://sites.google.com/view/seamlessnerf.

There are two main categories of image super-resolution algorithms: distortion oriented and perception oriented. Recent evidence shows that reconstruction accuracy and perceptual quality are typically in disagreement with each other. In this article, we present a new image super-resolution framework that is capable of striking a balance between distortion and perception. The core of our framework is a deep fusion network capable of generating a final high-resolution image by fusing a pair of deterministic and stochastic images using spatially varying weights. To make a single fusion model produce images with varying degrees of stochasticity, we further incorporate meta-learning into our fusion network. Once equipped with the kernel produced by a kernel prediction module, our meta fusion network is able to produce final images at any desired level of stochasticity. Experimental results indicate that our meta fusion network outperforms existing state-of-the-art SISR algorithms on widely used datasets, including PIRM-val, DIV2K-val, Set5, Set14, Urban100, Manga109, and B100. In addition, it is capable of producing high-resolution images that achieve low distortion and high perceptual quality simultaneously.

Point completion refers to completing the missing geometries of an object from incomplete observations. Mainstream methods predict the missing shapes by decoding a global feature learned from the input point cloud, which often leads to deficient results in preserving topology consistency and surface details. In this work, we present MEPCN, a point completion network that leverages emptiness in 3D shape space. Given a single depth scan, previous methods often encode the occupied partial shapes while ignoring the empty regions (e.g. holes) in depth maps. In contrast, we argue that these ‘emptiness’ clues indicate shape boundaries that can be used to improve topology representation and detail granularity on surfaces. Specifically, our ME-PCN encodes both the occupied point cloud and the neighboring ‘empty points’. It estimates coarse-grained but complete and reasonable surface points in the first stage, followed by a refinement stage to produce fine-grained surface details. Comprehensive experiments verify that our ME-PCN presents better qualitative and quantitative performance against the state-of-the-art. Besides, we further prove that our ‘emptiness’ design is lightweight and easy to embed in existing methods, which shows consistent effectiveness in improving the CD and EMD scores.

Single image superresolution has been a popular research topic in the last two decades and has recently received a new wave of interest due to deep neural networks. In this paper, we approach this problem from a different perspective. With respect to a downsampled low resolution image, we model a high resolution image as a combination of two components, a deterministic component and a stochastic component. The deterministic component can be recovered from the low-frequency signals in the downsampled image. The stochastic component, on the other hand, contains the signals that have little correlation with the low resolution image. We adopt two complementary methods for generating these two components. While generative adversarial networks are used for the stochastic component, deterministic component reconstruction is formulated as a regression problem solved using deep neural networks. Since the deterministic component exhibits clearer local orientations, we design novel loss functions tailored for such properties for training the deep regression network. These two methods are first applied to the entire input image to produce two distinct high-resolution images. Afterwards, these two images are fused together using another deep neural network that also performs local statistical rectification, which tries to make the local statistics of the fused image match the same local statistics of the groundtruth image. Quantitative results and a user study indicate that the proposed method outperforms existing state-of-the-art algorithms with a clear margin.

We introduce Tamp, an open source C++ library for reducing the space and time costs of deep neural network models. In particular, Tamp implements several recent works which use structured matrices to replace unstructured matrices which are often bottlenecks in neural networks.

We introduce music-driven video montage, a media format that offers a pleasant way to browse or summarize video clips collected from various occasions, including gatherings and adventures. In music-driven video montage, the music drives the composition of the video content. According to musical movement and beats, video clips are organized to form a montage that visually reflects the experiential properties of the music. Nonetheless, it takes enormous manual work and artistic expertise to create it. In this paper, we develop a framework for automatically generating music-driven video montages. The input is a set of video clips and a piece of background music. By analyzing the music and video content, our system extracts carefully designed temporal features from the input, and casts the synthesis problem as an optimization and solves the parameters through Markov Chain Monte Carlo sampling. The output is a video montage whose visual activities are cut and synchronized with the rhythm of the music, rendering a symphony of audio-visual resonance.