Hadar Averbuch-Elor

Jiaming Sun, Xi Chen, Qianqian Wang et al. · deepmind

We are witnessing an explosion of neural implicit representations in computer vision and graphics. Their applicability has recently expanded beyond tasks such as shape generation and image-based rendering to the fundamental problem of image-based 3D reconstruction. However, existing methods typically assume constrained 3D environments with constant illumination captured by a small set of roughly uniformly distributed cameras. We introduce a new method that enables efficient and accurate surface reconstruction from Internet photo collections in the presence of varying illumination. To achieve this, we propose a hybrid voxel- and surface-guided sampling technique that allows for more efficient ray sampling around surfaces and leads to significant improvements in reconstruction quality. Further, we present a new benchmark and protocol for evaluating reconstruction performance on such in-the-wild scenes. We perform extensive experiments, demonstrating that our approach surpasses both classical and neural reconstruction methods on a wide variety of metrics.

Ruojin Cai, Joseph Tung, Qianqian Wang et al. · deepmind

We consider the visual disambiguation task of determining whether a pair of visually similar images depict the same or distinct 3D surfaces (e.g., the same or opposite sides of a symmetric building). Illusory image matches, where two images observe distinct but visually similar 3D surfaces, can be challenging for humans to differentiate, and can also lead 3D reconstruction algorithms to produce erroneous results. We propose a learning-based approach to visual disambiguation, formulating it as a binary classification task on image pairs. To that end, we introduce a new dataset for this problem, Doppelgangers, which includes image pairs of similar structures with ground truth labels. We also design a network architecture that takes the spatial distribution of local keypoints and matches as input, allowing for better reasoning about both local and global cues. Our evaluation shows that our method can distinguish illusory matches in difficult cases, and can be integrated into SfM pipelines to produce correct, disambiguated 3D reconstructions. See our project page for our code, datasets, and more results: http://doppelgangers-3d.github.io/.

Haotong Lin, Qianqian Wang, Ruojin Cai et al. · deepmind

In this work, we aim to reconstruct a time-varying 3D model, capable of rendering photo-realistic renderings with independent control of viewpoint, illumination, and time, from Internet photos of large-scale landmarks. The core challenges are twofold. First, different types of temporal changes, such as illumination and changes to the underlying scene itself (such as replacing one graffiti artwork with another) are entangled together in the imagery. Second, scene-level temporal changes are often discrete and sporadic over time, rather than continuous. To tackle these problems, we propose a new scene representation equipped with a novel temporal step function encoding method that can model discrete scene-level content changes as piece-wise constant functions over time. Specifically, we represent the scene as a space-time radiance field with a per-image illumination embedding, where temporally-varying scene changes are encoded using a set of learned step functions. To facilitate our task of chronology reconstruction from Internet imagery, we also collect a new dataset of four scenes that exhibit various changes over time. We demonstrate that our method exhibits state-of-the-art view synthesis results on this dataset, while achieving independent control of viewpoint, time, and illumination.

Eric Ming Chen, Jin Sun, Apoorv Khandelwal et al. · deepmind, mit

How can one visually characterize people in a decade? In this work, we assemble the Faces Through Time dataset, which contains over a thousand portrait images from each decade, spanning the 1880s to the present day. Using our new dataset, we present a framework for resynthesizing portrait images across time, imagining how a portrait taken during a particular decade might have looked like, had it been taken in other decades. Our framework optimizes a family of per-decade generators that reveal subtle changes that differentiate decade--such as different hairstyles or makeup--while maintaining the identity of the input portrait. Experiments show that our method is more effective in resynthesizing portraits across time compared to state-of-the-art image-to-image translation methods, as well as attribute-based and language-guided portrait editing models. Our code and data will be available at https://facesthroughtime.github.io

Or Patashnik, Daniel Garibi, Idan Azuri et al.

Text-to-image models give rise to workflows which often begin with an exploration step, where users sift through a large collection of generated images. The global nature of the text-to-image generation process prevents users from narrowing their exploration to a particular object in the image. In this paper, we present a technique to generate a collection of images that depicts variations in the shape of a specific object, enabling an object-level shape exploration process. Creating plausible variations is challenging as it requires control over the shape of the generated object while respecting its semantics. A particular challenge when generating object variations is accurately localizing the manipulation applied over the object's shape. We introduce a prompt-mixing technique that switches between prompts along the denoising process to attain a variety of shape choices. To localize the image-space operation, we present two techniques that use the self-attention layers in conjunction with the cross-attention layers. Moreover, we show that these localization techniques are general and effective beyond the scope of generating object variations. Extensive results and comparisons demonstrate the effectiveness of our method in generating object variations, and the competence of our localization techniques.

Yuval Alaluf, Daniel Garibi, Or Patashnik et al.

Recent advancements in text-to-image generative models have demonstrated a remarkable ability to capture a deep semantic understanding of images. In this work, we leverage this semantic knowledge to transfer the visual appearance between objects that share similar semantics but may differ significantly in shape. To achieve this, we build upon the self-attention layers of these generative models and introduce a cross-image attention mechanism that implicitly establishes semantic correspondences across images. Specifically, given a pair of images -- one depicting the target structure and the other specifying the desired appearance -- our cross-image attention combines the queries corresponding to the structure image with the keys and values of the appearance image. This operation, when applied during the denoising process, leverages the established semantic correspondences to generate an image combining the desired structure and appearance. In addition, to improve the output image quality, we harness three mechanisms that either manipulate the noisy latent codes or the model's internal representations throughout the denoising process. Importantly, our approach is zero-shot, requiring no optimization or training. Experiments show that our method is effective across a wide range of object categories and is robust to variations in shape, size, and viewpoint between the two input images.

Noriyuki Kojima, Hadar Averbuch-Elor, Yoav Artzi

Key to tasks that require reasoning about natural language in visual contexts is grounding words and phrases to image regions. However, observing this grounding in contemporary models is complex, even if it is generally expected to take place if the task is addressed in a way that is conductive to generalization. We propose a framework to jointly study task performance and phrase grounding, and propose three benchmarks to study the relation between the two. Our results show that contemporary models demonstrate inconsistency between their ability to ground phrases and solve tasks. We show how this can be addressed through brute-force training on ground phrasing annotations, and analyze the dynamics it creates. Code and at available at https://github.com/lil-lab/phrase_grounding.

Guangzhao He, Rundong Luo, Wei-Chiu Ma et al.

Inverse graphics is a longstanding and highly underconstrained problem that seeks to reconstruct images as editable 3D scenes which can be rendered, relit, and manipulated. In this work, we investigate whether pretrained vision-language models (VLMs) can perform executable inverse graphics directly from a single image by reconstructing a scene as an editable Blender program, without relying on specialized 2D or 3D foundation models, differentiable rendering, or multi-view supervision. We introduce Staged Executable Inverse Graphics (SEIG), an agentic framework that reconstructs a 3D scene from a single image by progressively refining scene factors including geometry, materials, composition, and lighting directly in executable Blender code space. We evaluate our framework across diverse scenes using a range of reconstruction metrics spanning pixel-level, perceptual, and semantic fidelity. Our experiments show that staged reconstruction substantially improves reconstruction fidelity, highlighting the importance of task decomposition for executable inverse graphics with general-purpose VLMs. Finally, we showcase various downstream applications enabled by the reconstructed editable Blender scenes.

Etai Sella, Gal Fiebelman, Peter Hedman et al.

Large scale text-guided diffusion models have garnered significant attention due to their ability to synthesize diverse images that convey complex visual concepts. This generative power has more recently been leveraged to perform text-to-3D synthesis. In this work, we present a technique that harnesses the power of latent diffusion models for editing existing 3D objects. Our method takes oriented 2D images of a 3D object as input and learns a grid-based volumetric representation of it. To guide the volumetric representation to conform to a target text prompt, we follow unconditional text-to-3D methods and optimize a Score Distillation Sampling (SDS) loss. However, we observe that combining this diffusion-guided loss with an image-based regularization loss that encourages the representation not to deviate too strongly from the input object is challenging, as it requires achieving two conflicting goals while viewing only structure-and-appearance coupled 2D projections. Thus, we introduce a novel volumetric regularization loss that operates directly in 3D space, utilizing the explicit nature of our 3D representation to enforce correlation between the global structure of the original and edited object. Furthermore, we present a technique that optimizes cross-attention volumetric grids to refine the spatial extent of the edits. Extensive experiments and comparisons demonstrate the effectiveness of our approach in creating a myriad of edits which cannot be achieved by prior works.

Tamir Cohen, Leo Segre, Shay Shomer-Chai et al.

Reconstructing accurate 3D models of large-scale real-world scenes from unstructured, in-the-wild imagery remains a core challenge in computer vision, especially when the input views have little or no overlap. In such cases, existing reconstruction pipelines often produce multiple disconnected partial reconstructions or erroneously merge non-overlapping regions into overlapping geometry. In this work, we propose a framework that grounds each partial reconstruction to a complete reference model of the scene, enabling globally consistent alignment even in the absence of visual overlap. We obtain reference models from dense, geospatially accurate pseudo-synthetic renderings derived from Google Earth Studio. These renderings provide full scene coverage but differ substantially in appearance from real-world photographs. Our key insight is that, despite this significant domain gap, both domains share the same underlying scene semantics. We represent the reference model using 3D Gaussian Splatting, augmenting each Gaussian with semantic features, and formulate alignment as an inverse feature-based optimization scheme that estimates a global 6DoF pose and scale while keeping the reference model fixed. Furthermore, we introduce the WikiEarth dataset, which registers existing partial 3D reconstructions with pseudo-synthetic reference models. We demonstrate that our approach consistently improves global alignment when initialized with various classical and learning-based pipelines, while mitigating failure modes of state-of-the-art end-to-end models. All code and data will be released.

Morris Alper, Michael Fiman, Hadar Averbuch-Elor

Most humans use visual imagination to understand and reason about language, but models such as BERT reason about language using knowledge acquired during text-only pretraining. In this work, we investigate whether vision-and-language pretraining can improve performance on text-only tasks that involve implicit visual reasoning, focusing primarily on zero-shot probing methods. We propose a suite of visual language understanding (VLU) tasks for probing the visual reasoning abilities of text encoder models, as well as various non-visual natural language understanding (NLU) tasks for comparison. We also contribute a novel zero-shot knowledge probing method, Stroop probing, for applying models such as CLIP to text-only tasks without needing a prediction head such as the masked language modelling head of models like BERT. We show that SOTA multimodally trained text encoders outperform unimodally trained text encoders on the VLU tasks while being underperformed by them on the NLU tasks, lending new context to previously mixed results regarding the NLU capabilities of multimodal models. We conclude that exposure to images during pretraining affords inherent visual reasoning knowledge that is reflected in language-only tasks that require implicit visual reasoning. Our findings bear importance in the broader context of multimodal learning, providing principled guidelines for the choice of text encoders used in such contexts.

Morris Alper, Hadar Averbuch-Elor

While recent vision-and-language models (VLMs) like CLIP are a powerful tool for analyzing text and images in a shared semantic space, they do not explicitly model the hierarchical nature of the set of texts which may describe an image. Conversely, existing multimodal hierarchical representation learning methods require costly training from scratch, failing to leverage the knowledge encoded by state-of-the-art multimodal foundation models. In this work, we study the knowledge of existing foundation models, finding that they exhibit emergent understanding of visual-semantic hierarchies despite not being directly trained for this purpose. We propose the Radial Embedding (RE) framework for probing and optimizing hierarchical understanding, and contribute the HierarCaps dataset, a benchmark facilitating the study of hierarchical knowledge in image--text representations, constructed automatically via large language models. Our results show that foundation VLMs exhibit zero-shot hierarchical understanding, surpassing the performance of prior models explicitly designed for this purpose. Furthermore, we show that foundation models may be better aligned to hierarchical reasoning via a text-only fine-tuning phase, while retaining pretraining knowledge.

Morris Alper, Hadar Averbuch-Elor

Although the mapping between sound and meaning in human language is assumed to be largely arbitrary, research in cognitive science has shown that there are non-trivial correlations between particular sounds and meanings across languages and demographic groups, a phenomenon known as sound symbolism. Among the many dimensions of meaning, sound symbolism is particularly salient and well-demonstrated with regards to cross-modal associations between language and the visual domain. In this work, we address the question of whether sound symbolism is reflected in vision-and-language models such as CLIP and Stable Diffusion. Using zero-shot knowledge probing to investigate the inherent knowledge of these models, we find strong evidence that they do show this pattern, paralleling the well-known kiki-bouba effect in psycholinguistics. Our work provides a novel method for demonstrating sound symbolism and understanding its nature using computational tools. Our code will be made publicly available.

Etai Sella, Yoav Baron, Hadar Averbuch-Elor et al.

Recent diffusion-based image editing methods commonly rely on text or high-level instructions to guide the generation process, offering intuitive but coarse control. In contrast, we focus on explicit, prompt-free editing, where the user directly specifies the modification by cropping and pasting an object or sub-object into a chosen location within an image. This operation affords precise spatial and visual control, yet it introduces a fundamental challenge: preserving the identity of the pasted object while harmonizing it with its new context. We observe that attention maps in diffusion-based editing models inherently govern whether image regions are preserved or adapted for coherence. Building on this insight, we introduce LooseRoPE, a saliency-guided modulation of rotational positional encoding (RoPE) that loosens the positional constraints to continuously control the attention field of view. By relaxing RoPE in this manner, our method smoothly steers the model's focus between faithful preservation of the input image and coherent harmonization of the inserted object, enabling a balanced trade-off between identity retention and contextual blending. Our approach provides a flexible and intuitive framework for image editing, achieving seamless compositional results without textual descriptions or complex user input.

Etai Sella, Gal Fiebelman, Noam Atia et al.

We are witnessing rapid progress in automatically generating and manipulating 3D assets due to the availability of pretrained text-image diffusion models. However, time-consuming optimization procedures are required for synthesizing each sample, hindering their potential for democratizing 3D content creation. Conversely, 3D diffusion models now train on million-scale 3D datasets, yielding high-quality text-conditional 3D samples within seconds. In this work, we present Spice-E - a neural network that adds structural guidance to 3D diffusion models, extending their usage beyond text-conditional generation. At its core, our framework introduces a cross-entity attention mechanism that allows for multiple entities (in particular, paired input and guidance 3D shapes) to interact via their internal representations within the denoising network. We utilize this mechanism for learning task-specific structural priors in 3D diffusion models from auxiliary guidance shapes. We show that our approach supports a variety of applications, including 3D stylization, semantic shape editing and text-conditional abstraction-to-3D, which transforms primitive-based abstractions into highly-expressive shapes. Extensive experiments demonstrate that Spice-E achieves SOTA performance over these tasks while often being considerably faster than alternative methods. Importantly, this is accomplished without tailoring our approach for any specific task.

Morris Alper, Hadar Averbuch-Elor

Interactions between humans are diverse and context-dependent, but previous works have treated them as categorical, disregarding the heavy tail of possible interactions. We propose a new paradigm of learning human-human interactions as free text from a single still image, allowing for flexibility in modeling the unlimited space of situations and relationships between people. To overcome the absence of data labelled specifically for this task, we use knowledge distillation applied to synthetic caption data produced by a large language model without explicit supervision. We show that the pseudo-labels produced by this procedure can be used to train a captioning model to effectively understand human-human interactions in images, as measured by a variety of metrics that measure textual and semantic faithfulness and factual groundedness of our predictions. We further show that our approach outperforms SOTA image captioning and situation recognition models on this task. We will release our code and pseudo-labels along with Waldo and Wenda, a manually-curated test set for still image human-human interaction understanding.

Wenxuan Peng, Bharath Hariharan, Hadar Averbuch-Elor

Despite recent progress, text-to-image models still struggle to generate semantically diverse and compositionally accurate multi-person interaction scenes, often collapsing to repetitive layouts, stereotypical poses, and poorly grounded interactions. In this work, we bridge this gap by introducing a dual pose-image representation that brings person-centric structural priors into pretrained diffusion transformers. Our model jointly predicts a 2D pose visualization image and its corresponding RGB image, enabling structure and appearance to co-evolve during learning. At its core, a cross-modal alignment scheme binds text, pose, and image representations, ensuring consistent grounding across modalities. Furthermore, we design an iterative scene construction scheme, progressively generating complex multi-human interactions while effectively decomposing the overall generation complexity. Extensive experiments demonstrate that our method substantially improves prompt alignment and scene diversity in multi-person image generation.

Junhyeong Cho, Ruojin Cai, Hadar Averbuch-Elor

Many public buildings provide floorplans with a "you are here" indicator to help visitors orient themselves. Floorplan localization seeks to computationally replicate this capability by determining where visual observations were captured within a floorplan. However, existing methods typically assume controlled small-scale environments and precise vectorized floorplans, limiting their ability to operate in large-scale buildings and rasterized floorplans. In this work, we present an approach for performing floorplan localization in the wild by grounding the task in a reconstructed 3D representation of the scene. Given an unconstrained image collection, our method reconstructs a gravity-aligned 3D scene and projects it into a 2D density map that serves as a floorplan proxy. Floorplan localization is then formulated as aligning this proxy with the input floorplan via a 2D similarity transform. To bridge the appearance gap between density maps and architectural floorplans, we adapt a 2D foundation model to learn cross-modal correspondences, introducing a fine-tuning scheme that encourages semantically aligned matches while preserving structural consistency. Extensive experiments demonstrate substantial improvements over prior methods, including in extremely sparse settings with as little as a single input image. Our code and data will be publicly available.

Shai Krakovsky, Gal Fiebelman, Sagie Benaim et al.

Embedding a language field in a 3D representation enables richer semantic understanding of spatial environments by linking geometry with descriptive meaning. This allows for a more intuitive human-computer interaction, enabling querying or editing scenes using natural language, and could potentially improve tasks like scene retrieval, navigation, and multimodal reasoning. While such capabilities could be transformative, in particular for large-scale scenes, we find that recent feature distillation approaches cannot effectively learn over massive Internet data due to challenges in semantic feature misalignment and inefficiency in memory and runtime. To this end, we propose a novel approach to address these challenges. First, we introduce extremely low-dimensional semantic bottleneck features as part of the underlying 3D Gaussian representation. These are processed by rendering and passing them through a multi-resolution, feature-based, hash encoder. This significantly improves efficiency both in runtime and GPU memory. Second, we introduce an Attenuated Downsampler module and propose several regularizations addressing the semantic misalignment of ground truth 2D features. We evaluate our method on the in-the-wild HolyScenes dataset and demonstrate that it surpasses existing approaches in both performance and efficiency.

Hyunwoo Kim, Itai Lang, Hadar Averbuch-Elor et al.

We propose MeshOn, a method that finds physically and semantically realistic compositions of two input meshes. Given an accessory, a base mesh with a user-defined target region, and optional text strings for both meshes, MeshOn uses a multi-step optimization framework to realistically fit the meshes onto each other while preventing intersections. We initialize the shapes' rigid configuration via a structured alignment scheme using Vision-to-Language Models, which we then optimize using a combination of attractive geometric losses, and a physics-inspired barrier loss that prevents surface intersections. We then obtain a final deformation of the object, assisted by a diffusion prior. Our method successfully fits accessories of various materials over a breadth of target regions, and is designed to fit directly into existing digital artist workflows. We demonstrate the robustness and accuracy of our pipeline by comparing it with generative approaches and traditional registration algorithms.

Assaf Ben-Kish, Moran Yanuka, Morris Alper et al. · apple-ml

While recent years have seen rapid progress in image-conditioned text generation, image captioning still suffers from the fundamental issue of hallucinations, namely, the generation of spurious details that cannot be inferred from the given image. Existing methods largely use closed-vocabulary object lists to mitigate or evaluate hallucinations in image captioning, ignoring the long-tailed nature of hallucinations that occur in practice. To this end, we propose a framework for addressing hallucinations in image captioning in the open-vocabulary setting. Our framework includes a new benchmark, OpenCHAIR, that leverages generative foundation models to evaluate open-vocabulary object hallucinations for image captioning, surpassing the popular and similarly-sized CHAIR benchmark in both diversity and accuracy. Furthermore, to mitigate open-vocabulary hallucinations without using a closed object list, we propose MOCHa, an approach harnessing advancements in reinforcement learning. Our multi-objective reward function explicitly targets the trade-off between fidelity and adequacy in generations without requiring any strong supervision. MOCHa improves a large variety of image captioning models, as captured by our OpenCHAIR benchmark and other existing metrics. Code and models can be found at: https://github.com/assafbk/mocha_code

Keren Ganon, Morris Alper, Rachel Mikulinsky et al.

Buildings are a central feature of human culture and are increasingly being analyzed with computational methods. However, recent works on computational building understanding have largely focused on natural imagery of buildings, neglecting the fundamental element defining a building's structure -- its floorplan. Conversely, existing works on floorplan understanding are extremely limited in scope, often focusing on floorplans of a single semantic category and region (e.g. floorplans of apartments from a single country). In this work, we introduce WAFFLE, a novel multimodal floorplan understanding dataset of nearly 20K floorplan images and metadata curated from Internet data spanning diverse building types, locations, and data formats. By using a large language model and multimodal foundation models, we curate and extract semantic information from these images and their accompanying noisy metadata. We show that WAFFLE enables progress on new building understanding tasks, both discriminative and generative, which were not feasible using prior datasets. We will publicly release WAFFLE along with our code and trained models, providing the research community with a new foundation for learning the semantics of buildings.

Fengbei Liu, Sunwoo Kwak, Hao Phung et al.

Non-contrast chest CTs offer a rich opportunity for both conventional pulmonary and opportunistic extra-pulmonary screening. While Multi-Task Learning (MTL) can unify these diverse tasks, standard hard-parameter sharing approaches are often suboptimal for modeling distinct pathologies. We propose HyperCT, a framework that dynamically adapts a Vision Transformer backbone via a Hypernetwork. To ensure computational efficiency, we integrate Low-Rank Adaptation (LoRA), allowing the model to regress task-specific low-rank weight updates rather than full parameters. Validated on a large-scale dataset of radiological and cardiological tasks, \method{} outperforms various strong baselines, offering a unified, parameter-efficient solution for holistic patient assessment. Our code is available at https://github.com/lfb-1/HyperCT.

Rachel Mikulinsky, Morris Alper, Shai Gordin et al.

The cuneiform writing system served as the medium for transmitting knowledge in the ancient Near East for a period of over three thousand years. Cuneiform signs have a complex internal structure which is the subject of expert paleographic analysis, as variations in sign shapes bear witness to historical developments and transmission of writing and culture over time. However, prior automated techniques mostly treat sign types as categorical and do not explicitly model their highly varied internal configurations. In this work, we present an unsupervised approach for recovering the fine-grained internal configuration of cuneiform signs by leveraging powerful generative models and the appearance and structure of prototype font images as priors. Our approach, ProtoSnap, enforces structural consistency on matches found with deep image features to estimate the diverse configurations of cuneiform characters, snapping a skeleton-based template to photographed cuneiform signs. We provide a new benchmark of expert annotations and evaluate our method on this task. Our evaluation shows that our approach succeeds in aligning prototype skeletons to a wide variety of cuneiform signs. Moreover, we show that conditioning on structures produced by our method allows for generating synthetic data with correct structural configurations, significantly boosting the performance of cuneiform sign recognition beyond existing techniques, in particular over rare signs. Our code, data, and trained models are available at the project page: https://tau-vailab.github.io/ProtoSnap/

Shahar Zuler, Gal Lifshitz, Hadar Averbuch-Elor et al.

Accurate motion estimation in cardiac computed tomography (CT) imaging is critical for assessing cardiac function and surgical planning. Data-driven methods have become the standard approach for dense motion estimation, but they rely on vast amounts of labeled data with dense ground-truth (GT) motion annotations, which are often unfeasible to obtain. To address this limitation, we present a novel approach that synthesizes realistically looking pairs of cardiac CT frames enriched with dense 3D flow field annotations. Our method leverages a conditional Variational Autoencoder (CVAE), which incorporates a novel multi-scale feature conditioning mechanism and is trained to generate 3D flow fields conditioned on a single CT frame. By applying the generated flow field to warp the given frame, we create pairs of frames that simulate realistic myocardium deformations across the cardiac cycle. These pairs serve as fully annotated data samples, providing optical flow GT annotations. Our data generation pipeline could enable the training and validation of more complex and accurate myocardium motion models, allowing for substantially reducing reliance on manual annotations. Our code, along with animated generated samples and additional material, is available on our project page: https://shaharzuler.github.io/GenerativeCardiacMotion_Page.

Ruojin Cai, Guandao Yang, Hadar Averbuch-Elor et al.

In this work, we propose a novel technique to generate shapes from point cloud data. A point cloud can be viewed as samples from a distribution of 3D points whose density is concentrated near the surface of the shape. Point cloud generation thus amounts to moving randomly sampled points to high-density areas. We generate point clouds by performing stochastic gradient ascent on an unnormalized probability density, thereby moving sampled points toward the high-likelihood regions. Our model directly predicts the gradient of the log density field and can be trained with a simple objective adapted from score-based generative models. We show that our method can reach state-of-the-art performance for point cloud auto-encoding and generation, while also allowing for extraction of a high-quality implicit surface. Code is available at https://github.com/RuojinCai/ShapeGF.

Daniel Garibi, Or Patashnik, Andrey Voynov et al.

Recent advancements in text-guided diffusion models have unlocked powerful image manipulation capabilities. However, applying these methods to real images necessitates the inversion of the images into the domain of the pretrained diffusion model. Achieving faithful inversion remains a challenge, particularly for more recent models trained to generate images with a small number of denoising steps. In this work, we introduce an inversion method with a high quality-to-operation ratio, enhancing reconstruction accuracy without increasing the number of operations. Building on reversing the diffusion sampling process, our method employs an iterative renoising mechanism at each inversion sampling step. This mechanism refines the approximation of a predicted point along the forward diffusion trajectory, by iteratively applying the pretrained diffusion model, and averaging these predictions. We evaluate the performance of our ReNoise technique using various sampling algorithms and models, including recent accelerated diffusion models. Through comprehensive evaluations and comparisons, we show its effectiveness in terms of both accuracy and speed. Furthermore, we confirm that our method preserves editability by demonstrating text-driven image editing on real images.

Yuan Li, Yuanbo Xiangli, Hadar Averbuch-Elor et al.

Internet photo collections exhibit an extremely long-tailed distribution: a few famous landmarks are densely photographed and easily reconstructed in 3D, while most real-world sites are represented with sparse, noisy, uneven imagery beyond the capabilities of both classical and learned 3D methods. We believe that tackling this long-tail regime represents one of the next frontiers for 3D foundation models. Although reliable ground-truth 3D supervision from sparse scenes is challenging to acquire, we observe that it can be effectively simulated by sampling sparse subsets from well-reconstructed Internet landmarks. To this end, we introduce MegaDepth-X, a large dataset of 3D reconstructions with clean, dense depth, together with a strategy for sampling sets of training images that mimic camera distributions in long-tail scenes. Finetuning 3D foundation models with these components yields robust reconstructions under extreme sparsity, and also enables more reliable reconstruction in symmetric and repetitive scenes, while preserving generalization to standard, dense 3D benchmark datasets.

Hao Phung, Hadar Averbuch-Elor

Reconstructing a structured vector-graphics representation from a rasterized floorplan image is typically an important prerequisite for computational tasks involving floorplans such as automated understanding or CAD workflows. However, existing techniques struggle in faithfully generating the structure and semantics conveyed by complex floorplans that depict large indoor spaces with many rooms and a varying numbers of polygon corners. To this end, we propose Raster2Seq, framing floorplan reconstruction as a sequence-to-sequence task in which floorplan elements--such as rooms, windows, and doors--are represented as labeled polygon sequences that jointly encode geometry and semantics. Our approach introduces an autoregressive decoder that learns to predict the next corner conditioned on image features and previously generated corners using guidance from learnable anchors. These anchors represent spatial coordinates in image space, hence allowing for effectively directing the attention mechanism to focus on informative image regions. By embracing the autoregressive mechanism, our method offers flexibility in the output format, enabling for efficiently handling complex floorplans with numerous rooms and diverse polygon structures. Our method achieves state-of-the-art performance on standard benchmarks such as Structure3D, CubiCasa5K, and Raster2Graph, while also demonstrating strong generalization to more challenging datasets like WAFFLE, which contain diverse room structures and complex geometric variations.

Moran Yanuka, Morris Alper, Hadar Averbuch-Elor et al. · apple-ml

Web-scale training on paired text-image data is becoming increasingly central to multimodal learning, but is challenged by the highly noisy nature of datasets in the wild. Standard data filtering approaches succeed in removing mismatched text-image pairs, but permit semantically related but highly abstract or subjective text. These approaches lack the fine-grained ability to isolate the most concrete samples that provide the strongest signal for learning in a noisy dataset. In this work, we propose a new metric, image caption concreteness, that evaluates caption text without an image reference to measure its concreteness and relevancy for use in multimodal learning. Our approach leverages strong foundation models for measuring visual-semantic information loss in multimodal representations. We demonstrate that this strongly correlates with human evaluation of concreteness in both single-word and sentence-level texts. Moreover, we show that curation using ICC complements existing approaches: It succeeds in selecting the highest quality samples from multimodal web-scale datasets to allow for efficient training in resource-constrained settings.

Etai Sella, Hao Phung, Nitay Amiel et al.

Text-based 2D image editing models have recently reached an impressive level of maturity, motivating a growing body of work that heavily depends on these models to drive 3D edits. While effective for appearance-based modifications, such 2D-centric 3D editing pipelines often struggle with fine-grained 3D editing, where localized structural changes must be applied while strictly preserving an object's overall identity. To address this limitation, we propose Prox-E, a training-free framework that enables fine-grained 3D control through an explicit, primitive-based geometric abstraction. Our framework first abstracts an input 3D shape into a compact set of geometric primitives. A pretrained vision-language model (VLM) then edits this abstraction to specify primitive-level changes. These structural edits are subsequently used to guide a 3D generative model, enabling fine-grained, localized modifications while preserving unchanged regions of the original shape. Through extensive experiments, we demonstrate that our method consistently balances identity preservation, shape quality, and instruction fidelity more effectively than various existing approaches, including 2D-based 3D editors and training-based methods.

Gal Fiebelman, Tamir Cohen, Ayellet Morgenstern et al.

The emergence of neural representations has revolutionized our means for digitally viewing a wide range of 3D scenes, enabling the synthesis of photorealistic images rendered from novel views. Recently, several techniques have been proposed for connecting these low-level representations with the high-level semantics understanding embodied within the scene. These methods elevate the rich semantic understanding from 2D imagery to 3D representations, distilling high-dimensional spatial features onto 3D space. In our work, we are interested in connecting language with a dynamic modeling of the world. We show how to lift spatio-temporal features to a 4D representation based on 3D Gaussian Splatting. This enables an interactive interface where the user can spatiotemporally localize events in the video from text prompts. We demonstrate our system on public 3D video datasets of people and animals performing various actions.

Hana Bezalel, Dotan Ankri, Ruojin Cai et al.

We present a technique and benchmark dataset for estimating the relative 3D orientation between a pair of Internet images captured in an extreme setting, where the images have limited or non-overlapping field of views. Prior work targeting extreme rotation estimation assume constrained 3D environments and emulate perspective images by cropping regions from panoramic views. However, real images captured in the wild are highly diverse, exhibiting variation in both appearance and camera intrinsics. In this work, we propose a Transformer-based method for estimating relative rotations in extreme real-world settings, and contribute the ExtremeLandmarkPairs dataset, assembled from scene-level Internet photo collections. Our evaluation demonstrates that our approach succeeds in estimating the relative rotations in a wide variety of extreme-view Internet image pairs, outperforming various baselines, including dedicated rotation estimation techniques and contemporary 3D reconstruction methods.

Chen Dudai, Morris Alper, Hana Bezalel et al.

Internet image collections containing photos captured by crowds of photographers show promise for enabling digital exploration of large-scale tourist landmarks. However, prior works focus primarily on geometric reconstruction and visualization, neglecting the key role of language in providing a semantic interface for navigation and fine-grained understanding. In constrained 3D domains, recent methods have leveraged vision-and-language models as a strong prior of 2D visual semantics. While these models display an excellent understanding of broad visual semantics, they struggle with unconstrained photo collections depicting such tourist landmarks, as they lack expert knowledge of the architectural domain. In this work, we present a localization system that connects neural representations of scenes depicting large-scale landmarks with text describing a semantic region within the scene, by harnessing the power of SOTA vision-and-language models with adaptations for understanding landmark scene semantics. To bolster such models with fine-grained knowledge, we leverage large-scale Internet data containing images of similar landmarks along with weakly-related textual information. Our approach is built upon the premise that images physically grounded in space can provide a powerful supervision signal for localizing new concepts, whose semantics may be unlocked from Internet textual metadata with large language models. We use correspondences between views of scenes to bootstrap spatial understanding of these semantics, providing guidance for 3D-compatible segmentation that ultimately lifts to a volumetric scene representation. Our results show that HaLo-NeRF can accurately localize a variety of semantic concepts related to architectural landmarks, surpassing the results of other 3D models as well as strong 2D segmentation baselines. Our project page is at https://tau-vailab.github.io/HaLo-NeRF/.

Gal Fiebelman, Hadar Averbuch-Elor, Sagie Benaim

3D Gaussian Splatting has recently enabled fast and photorealistic reconstruction of static 3D scenes. However, introducing dynamic elements that interact naturally with such static scenes remains challenging. Accordingly, we present a novel hybrid framework that combines Gaussian-particle representations for incorporating physically-based global weather effects into static 3D Gaussian Splatting scenes, correctly handling the interactions of dynamic elements with the static scene. We follow a three-stage process: we first map static 3D Gaussians to a particle-based representation. We then introduce dynamic particles and simulate their motion using the Material Point Method (MPM). Finally, we map the simulated particles back to the Gaussian domain while introducing appearance parameters tailored for specific effects. To correctly handle the interactions of dynamic elements with the static scene, we introduce specialized collision handling techniques. Our approach supports a variety of weather effects, including snowfall, rainfall, fog, and sandstorms, and can also support falling objects, all with physically plausible motion and appearance. Experiments demonstrate that our method significantly outperforms existing approaches in both visual quality and physical realism.

Shay Shomer Chai, Wenxuan Peng, Bharath Hariharan et al.

Text-to-image generation has recently seen remarkable success, granting users with the ability to create high-quality images through the use of text. However, contemporary methods face challenges in capturing the precise semantics conveyed by complex multi-object prompts. Consequently, many works have sought to mitigate such semantic misalignments, typically via inference-time schemes that modify the attention layers of the denoising networks. However, prior work has mostly utilized coarse metrics, such as the cosine similarity between text and image CLIP embeddings, or human evaluations, which are challenging to conduct on a larger-scale. In this work, we perform a case study on colors -- a fundamental attribute commonly associated with objects in text prompts, which offer a rich test bed for rigorous evaluation. Our analysis reveals that pretrained models struggle to generate images that faithfully reflect multiple color attributes-far more so than with single-color prompts-and that neither inference-time techniques nor existing editing methods reliably resolve these semantic misalignments. Accordingly, we introduce a dedicated image editing technique, mitigating the issue of multi-object semantic alignment for prompts containing multiple colors. We demonstrate that our approach significantly boosts performance over a wide range of metrics, considering images generated by various text-to-image diffusion-based techniques.

Yuval Grader, Hadar Averbuch-Elor

Floorplans provide a compact representation of the building's structure, revealing not only layout information but also detailed semantics such as the locations of windows and doors. However, contemporary floorplan localization techniques mostly focus on matching depth-based structural cues, ignoring the rich semantics communicated within floorplans. In this work, we introduce a semantic-aware localization framework that jointly estimates depth and semantic rays, consolidating over both for predicting a structural-semantic probability volume. Our probability volume is constructed in a coarse-to-fine manner: We first sample a small set of rays to obtain an initial low-resolution probability volume. We then refine these probabilities by performing a denser sampling only in high-probability regions and process the refined values for predicting a 2D location and orientation angle. We conduct an evaluation on two standard floorplan localization benchmarks. Our experiments demonstrate that our approach substantially outperforms state-of-the-art methods, achieving significant improvements in recall metrics compared to prior works. Moreover, we show that our framework can easily incorporate additional metadata such as room labels, enabling additional gains in both accuracy and efficiency.

Morris Alper, David Novotny, Filippos Kokkinos et al.

Despite recent advances in sparse novel view synthesis (NVS) applied to object-centric scenes, scene-level NVS remains a challenge. A central issue is the lack of available clean multi-view training data, beyond manually curated datasets with limited diversity, camera variation, or licensing issues. On the other hand, an abundance of diverse and permissively-licensed data exists in the wild, consisting of scenes with varying appearances (illuminations, transient occlusions, etc.) from sources such as tourist photos. To this end, we present WildCAT3D, a framework for generating novel views of scenes learned from diverse 2D scene image data captured in the wild. We unlock training on these data sources by explicitly modeling global appearance conditions in images, extending the state-of-the-art multi-view diffusion paradigm to learn from scene views of varying appearances. Our trained model generalizes to new scenes at inference time, enabling the generation of multiple consistent novel views. WildCAT3D provides state-of-the-art results on single-view NVS in object- and scene-level settings, while training on strictly less data sources than prior methods. Additionally, it enables novel applications by providing global appearance control during generation.

Etai Sella, Yanir Kleiman, Hadar Averbuch-Elor

Despite rapid advancements in the capabilities of generative models, pretrained text-to-image models still struggle in capturing the semantics conveyed by complex prompts that compound multiple objects and instance-level attributes. Consequently, we are witnessing growing interests in integrating additional structural constraints, typically in the form of coarse bounding boxes, to better guide the generation process in such challenging cases. In this work, we take the idea of structural guidance a step further by making the observation that contemporary image generation models can directly provide a plausible fine-grained structural initialization. We propose a technique that couples this image-based structural guidance with LLM-based instance-level instructions, yielding output images that adhere to all parts of the text prompt, including object counts, instance-level attributes, and spatial relations between instances.

Shahaf Pruss, Morris Alper, Hadar Averbuch-Elor

Images depicting complex, dynamic scenes are challenging to parse automatically, requiring both high-level comprehension of the overall situation and fine-grained identification of participating entities and their interactions. Current approaches use distinct methods tailored to sub-tasks such as Situation Recognition and detection of Human-Human and Human-Object Interactions. However, recent advances in image understanding have often leveraged web-scale vision-language (V&L) representations to obviate task-specific engineering. In this work, we propose a framework for dynamic scene understanding tasks by leveraging knowledge from modern, frozen V&L representations. By framing these tasks in a generic manner - as predicting and parsing structured text, or by directly concatenating representations to the input of existing models - we achieve state-of-the-art results while using a minimal number of trainable parameters relative to existing approaches. Moreover, our analysis of dynamic knowledge of these representations shows that recent, more powerful representations effectively encode dynamic scene semantics, making this approach newly possible.

Yiwen Zhang, Joseph Tung, Ruojin Cai et al.

3D foundation models (3DFMs) have recently transformed 3D vision, enabling joint prediction of depths, poses, and point maps directly from images. Yet their ability to reason under extreme, non-overlapping views remains largely unexplored. In this work, we study their internal representations and find that 3DFMs exhibit an emergent understanding of extreme-view geometry, despite never being trained for such conditions. To further enhance these capabilities, we introduce a lightweight alignment scheme that refines their internal 3D representation by tuning only a small subset of backbone bias terms, leaving all decoder heads frozen. This targeted adaptation substantially improves relative pose estimation under extreme viewpoints without degrading per-image depth or point quality. Additionally, we contribute MegaUnScene, a new benchmark of Internet scenes unseen by existing 3DFMs, with dedicated test splits for both relative pose estimation and dense 3D reconstruction. All code and data will be released.

Etai Sella, Noam Atia, Ron Mokady et al.

Natural language offers a highly intuitive interface for enabling localized fine-grained edits of 3D shapes. However, prior works face challenges in preserving global coherence while locally modifying the input 3D shape. In this work, we introduce an inpainting-based framework for editing shapes represented as point clouds. Our approach leverages foundation 3D diffusion models for achieving localized shape edits, adding structural guidance in the form of a partial conditional shape, ensuring that other regions correctly preserve the shape's identity. Furthermore, to encourage identity preservation also within the local edited region, we propose an inference-time coordinate blending algorithm which balances reconstruction of the full shape with inpainting at a progression of noise levels during the inference process. Our coordinate blending algorithm seamlessly blends the original shape with its edited version, enabling a fine-grained editing of 3D shapes, all while circumventing the need for computationally expensive and often inaccurate inversion. Extensive experiments show that our method outperforms alternative techniques across a wide range of metrics that evaluate both fidelity to the original shape and also adherence to the textual description.

Yotam Elor, Hadar Averbuch-Elor

Balancing the data before training a classifier is a popular technique to address the challenges of imbalanced binary classification in tabular data. Balancing is commonly achieved by duplication of minority samples or by generation of synthetic minority samples. While it is well known that balancing affects each classifier differently, most prior empirical studies did not include strong state-of-the-art (SOTA) classifiers as baselines. In this work, we are interested in understanding whether balancing is beneficial, particularly in the context of SOTA classifiers. Thus, we conduct extensive experiments considering three SOTA classifiers along the weaker learners used in previous investigations. Additionally, we carefully discern proper metrics, consistent and non-consistent algorithms and hyper-parameter selection methods and show that these have a significant impact on prediction quality and on the effectiveness of balancing. Our results support the known utility of balancing for weak classifiers. However, we find that balancing does not improve prediction performance for the strong ones. We further identify several other scenarios for which balancing is effective and observe that prior studies demonstrated the utility of balancing by focusing on these settings.

Claire Yuqing Cui, Apoorv Khandelwal, Yoav Artzi et al.

We present a task and benchmark dataset for person-centric visual grounding, the problem of linking between people named in a caption and people pictured in an image. In contrast to prior work in visual grounding, which is predominantly object-based, our new task masks out the names of people in captions in order to encourage methods trained on such image-caption pairs to focus on contextual cues (such as rich interactions between multiple people), rather than learning associations between names and appearances. To facilitate this task, we introduce a new dataset, Who's Waldo, mined automatically from image-caption data on Wikimedia Commons. We propose a Transformer-based method that outperforms several strong baselines on this task, and are releasing our data to the research community to spur work on contextual models that consider both vision and language.

Xiaoshi Wu, Hadar Averbuch-Elor, Jin Sun et al.

The abundance and richness of Internet photos of landmarks and cities has led to significant progress in 3D vision over the past two decades, including automated 3D reconstructions of the world's landmarks from tourist photos. However, a major source of information available for these 3D-augmented collections---namely language, e.g., from image captions---has been virtually untapped. In this work, we present WikiScenes, a new, large-scale dataset of landmark photo collections that contains descriptive text in the form of captions and hierarchical category names. WikiScenes forms a new testbed for multimodal reasoning involving images, text, and 3D geometry. We demonstrate the utility of WikiScenes for learning semantic concepts over images and 3D models. Our weakly-supervised framework connects images, 3D structure, and semantics---utilizing the strong constraints provided by 3D geometry---to associate semantic concepts to image pixels and 3D points.

Ruojin Cai, Bharath Hariharan, Noah Snavely et al.

We present a technique for estimating the relative 3D rotation of an RGB image pair in an extreme setting, where the images have little or no overlap. We observe that, even when images do not overlap, there may be rich hidden cues as to their geometric relationship, such as light source directions, vanishing points, and symmetries present in the scene. We propose a network design that can automatically learn such implicit cues by comparing all pairs of points between the two input images. Our method therefore constructs dense feature correlation volumes and processes these to predict relative 3D rotations. Our predictions are formed over a fine-grained discretization of rotations, bypassing difficulties associated with regressing 3D rotations. We demonstrate our approach on a large variety of extreme RGB image pairs, including indoor and outdoor images captured under different lighting conditions and geographic locations. Our evaluation shows that our model can successfully estimate relative rotations among non-overlapping images without compromising performance over overlapping image pairs.

Or Perel, Oron Anschel, Omri Ben-Eliezer et al.

Nowadays, as cameras are rapidly adopted in our daily routine, images of documents are becoming both abundant and prevalent. Unlike natural images that capture physical objects, document-images contain a significant amount of text with critical semantics and complicated layouts. In this work, we devise a generic unsupervised technique to learn multimodal affinities between textual entities in a document-image, considering their visual style, the content of their underlying text and their geometric context within the image. We then use these learned affinities to automatically cluster the textual entities in the image into different semantic groups. The core of our approach is a deep optimization scheme dedicated for an image provided by the user that detects and leverages reliable pairwise connections in the multimodal representation of the textual elements in order to properly learn the affinities. We show that our technique can operate on highly varying images spanning a wide range of documents and demonstrate its applicability for various editing operations manipulating the content, appearance and geometry of the image.

Margot Hanley, Apoorv Khandelwal, Hadar Averbuch-Elor et al.

Important ethical concerns arising from computer vision datasets of people have been receiving significant attention, and a number of datasets have been withdrawn as a result. To meet the academic need for people-centric datasets, we propose an analytical framework to guide ethical evaluation of existing datasets and to serve future dataset creators in avoiding missteps. Our work is informed by a review and analysis of prior works and highlights where such ethical challenges arise.

Jin Sun, Hadar Averbuch-Elor, Qianqian Wang et al.

Predicting where people can walk in a scene is important for many tasks, including autonomous driving systems and human behavior analysis. Yet learning a computational model for this purpose is challenging due to semantic ambiguity and a lack of labeled data: current datasets only tell you where people are, not where they could be. We tackle this problem by leveraging information from existing datasets, without additional labeling. We first augment the set of valid, labeled walkable regions by propagating person observations between images, utilizing 3D information to create what we call hidden footprints. However, this augmented data is still sparse. We devise a training strategy designed for such sparse labels, combining a class-balanced classification loss with a contextual adversarial loss. Using this strategy, we demonstrate a model that learns to predict a walkability map from a single image. We evaluate our model on the Waymo and Cityscapes datasets, demonstrating superior performance compared to baselines and state-of-the-art models.

Anna Darzi, Itai Lang, Ashutosh Taklikar et al.

As image generation techniques mature, there is a growing interest in explainable representations that are easy to understand and intuitive to manipulate. In this work, we turn to co-occurrence statistics, which have long been used for texture analysis, to learn a controllable texture synthesis model. We propose a fully convolutional generative adversarial network, conditioned locally on co-occurrence statistics, to generate arbitrarily large images while having local, interpretable control over the texture appearance. To encourage fidelity to the input condition, we introduce a novel differentiable co-occurrence loss that is integrated seamlessly into our framework in an end-to-end fashion. We demonstrate that our solution offers a stable, intuitive and interpretable latent representation for texture synthesis, which can be used to generate a smooth texture morph between different textures. We further show an interactive texture tool that allows a user to adjust local characteristics of the synthesized texture image using the co-occurrence values directly.