Scene generation with 3D assets presents a complex challenge, requiring both high-level semantic understanding and low-level geometric reasoning. While Multimodal Large Language Models (MLLMs) excel at semantic tasks, their application to 3D scene generation is hindered by their limited grounding on 3D geometry. In this paper, we investigate how to best work with MLLMs in an object placement task. Towards this goal, we introduce a novel framework, FirePlace, that applies existing MLLMs in (1) 3D geometric reasoning and the extraction of relevant geometric details from the 3D scene, (2) constructing and solving geometric constraints on the extracted low-level geometry, and (3) pruning for final placements that conform to common sense. By combining geometric reasoning with real-world understanding of MLLMs, our method can propose object placements that satisfy both geometric constraints as well as high-level semantic common-sense considerations. Our experiments show that these capabilities allow our method to place objects more effectively in complex scenes with intricate geometry, surpassing the quality of prior work.

While image-text representation learning has become very popular in recent years, existing models tend to lack spatial awareness and have limited direct applicability for dense understanding tasks. For this reason, self-supervised image-only pretraining is still the go-to method for many dense vision applications (e.g. depth estimation, semantic segmentation), despite the lack of explicit supervisory signals. In this paper, we close this gap between image-text and self-supervised learning, by proposing a novel general-purpose image-text model, which can be effectively used off the shelf for dense and global vision tasks. Our method, which we refer to as Text-Image Pretraining with Spatial awareness (TIPS), leverages two simple and effective insights. First, on textual supervision: we reveal that replacing noisy web image captions by synthetically generated textual descriptions boosts dense understanding performance significantly, due to a much richer signal for learning spatially aware representations. We propose an adapted training method that combines noisy and synthetic captions, resulting in improvements across both dense and global understanding tasks. Second, on the learning technique: we propose to combine contrastive image-text learning with self-supervised masked image modeling, to encourage spatial coherence, unlocking substantial enhancements for downstream applications. Building on these two ideas, we scale our model using the transformer architecture, trained on a curated set of public images. Our experiments are conducted on $8$ tasks involving $16$ datasets in total, demonstrating strong off-the-shelf performance on both dense and global understanding, for several image-only and image-text tasks. Code and models are released at https://github.com/google-deepmind/tips .

Finding localized correspondences across different images of the same object is crucial to understand its geometry. In recent years, this problem has seen remarkable progress with the advent of deep learning-based local image features and learnable matchers. Still, learnable matchers often underperform when there exists only small regions of co-visibility between image pairs (i.e. wide camera baselines). To address this problem, we leverage recent progress in coarse single-view geometry estimation methods. We propose LFM-3D, a Learnable Feature Matching framework that uses models based on graph neural networks and enhances their capabilities by integrating noisy, estimated 3D signals to boost correspondence estimation. When integrating 3D signals into the matcher model, we show that a suitable positional encoding is critical to effectively make use of the low-dimensional 3D information. We experiment with two different 3D signals - normalized object coordinates and monocular depth estimates - and evaluate our method on large-scale (synthetic and real) datasets containing object-centric image pairs across wide baselines. We observe strong feature matching improvements compared to 2D-only methods, with up to +6% total recall and +28% precision at fixed recall. Additionally, we demonstrate that the resulting improved correspondences lead to much higher relative posing accuracy for in-the-wild image pairs - up to 8.6% compared to the 2D-only approach.

From content moderation to wildlife conservation, the number of applications that require models to recognize nuanced or subjective visual concepts is growing. Traditionally, developing classifiers for such concepts requires substantial manual effort measured in hours, days, or even months to identify and annotate data needed for training. Even with recently proposed Agile Modeling techniques, which enable rapid bootstrapping of image classifiers, users are still required to spend 30 minutes or more of monotonous, repetitive data labeling just to train a single classifier. Drawing on Fiske's Cognitive Miser theory, we propose a new framework that alleviates manual effort by replacing human labeling with natural language interactions, reducing the total effort required to define a concept by an order of magnitude: from labeling 2,000 images to only 100 plus some natural language interactions. Our framework leverages recent advances in foundation models, both large language models and vision-language models, to carve out the concept space through conversation and by automatically labeling training data points. Most importantly, our framework eliminates the need for crowd-sourced annotations. Moreover, our framework ultimately produces lightweight classification models that are deploy-able in cost-sensitive scenarios. Across 15 subjective concepts and across 2 public image classification datasets, our trained models outperform traditional Agile Modeling as well as state-of-the-art zero-shot classification models like ALIGN, CLIP, CuPL, and large visual question answering models like PaLI-X.

We propose Encyclopedic-VQA, a large scale visual question answering (VQA) dataset featuring visual questions about detailed properties of fine-grained categories and instances. It contains 221k unique question+answer pairs each matched with (up to) 5 images, resulting in a total of 1M VQA samples. Moreover, our dataset comes with a controlled knowledge base derived from Wikipedia, marking the evidence to support each answer. Empirically, we show that our dataset poses a hard challenge for large vision+language models as they perform poorly on our dataset: PaLI [12] is state-of-the-art on OK-VQA [35], yet it only achieves 13.0% accuracy on our dataset. Moreover, we experimentally show that progress on answering our encyclopedic questions can be achieved by augmenting large models with a mechanism that retrieves relevant information from the knowledge base. An oracle experiment with perfect retrieval achieves 87.0% accuracy on the single-hop portion of our dataset, and an automatic retrieval-augmented prototype yields 48.8%. We believe that our dataset1enables future research on retrieval-augmented vision+language models.

Recent advances in neural reconstruction enable high-quality 3D object reconstruction from casually captured image collections. Current techniques mostly analyze their progress on relatively simple image collections where SfM techniques can provide ground-truth (GT) camera poses. We note that SfM techniques tend to fail on in-the-wild image collections such as image search results with varying backgrounds and illuminations. To enable systematic research progress on 3D reconstruction from casual image captures, we propose `NAVI': a new dataset of category-agnostic image collections of objects with high-quality 3D scans along with per-image 2D-3D alignments providing near-perfect GT camera parameters. These 2D-3D alignments allow us to extract accurate derivative annotations such as dense pixel correspondences, depth and segmentation maps. We demonstrate the use of NAVI image collections on different problem settings and show that NAVI enables more thorough evaluations that were not possible with existing datasets. We believe NAVI is beneficial for systematic research progress on 3D reconstruction and correspondence estimation.

We present a method that synthesizes novel views of complex scenes by interpolating a sparse set of nearby views. The core of our method is a network architecture that includes a multilayer perceptron and a ray transformer that estimates radiance and volume density at continuous 5D locations (3D spatial locations and 2D viewing directions), drawing appearance information on the fly from multiple source views. By drawing on source views at render time, our method hearkens back to classic work on image-based rendering (IBR), and allows us to render high-resolution imagery. Unlike neural scene representation work that optimizes per-scene functions for rendering, we learn a generic view interpolation function that generalizes to novel scenes. We render images using classic volume rendering, which is fully differentiable and allows us to train using only multi-view posed images as supervision. Experiments show that our method outperforms recent novel view synthesis methods that also seek to generalize to novel scenes. Further, if fine-tuned on each scene, our method is competitive with state-of-the-art single-scene neural rendering methods.1

Most deep architectures for image classification-even those that are trained to classify a large number of diverse categories-learn shared image representations with a single model. Intuitively, however, categories that are more similar should share more information than those that are very different. While hierarchical deep networks address this problem by learning separate features for subsets of related categories, current implementations require simplified models using fixed architectures specified via heuristic clustering methods. Instead, we propose Blockout, a method for regularization and model selection that simultaneously learns both the model architecture and parameters. A generalization of Dropout, our approach gives a novel parametrization of hierarchical architectures that allows for structure learning via back-propagation. To demonstrate its utility, we evaluate Blockout on the CIFAR and Image Net datasets, demonstrating improved classification accuracy, better regularization performance, faster training, and the clear emergence of hierarchical network structures.

The popular Richardson-Lucy (RL) image deconvolution algorithm often produces undesirable ringing artifacts. In this paper, we propose a novel Dual Range Deringing (DRD) algorithm to address this problem. As a post-deconvolution scheme, the proposed approach follows RL deconvolution and removes ringing artifacts by utilizing information from both the input blurred image and the RL-deblurred image. DRD first marks smooth regions in the input blurred image that are likely to be subjected to ringing artifacts far away from any strong edge. It then identifies short-range ringing artifacts from the regions that surround strong edges in the RL-deblurred image. Once marked, both long- and short-range ringing artifacts are then suppressed by an edge-preserving deringing filter. We demonstrate the effectiveness of this procedure by performing experiments on a set of images blurred with various Point Spread Functions (PSFs). We compare DRD with state-of-the-art non-blind deconvolution algorithms and show that our results are virtually free of ringing artifacts with only minor detail losses. Moreover, DRD consists of computationally efficient local operations and is suitable for parallelization on modern GPUs.

This paper presents a method for movie genre categorization of movie trailers, based on scene categorization. We view our approach as a step forward from using only low-level visual feature cues, towards the eventual goal of high-level seman- tic understanding of feature films. Our approach decom- poses each trailer into a collection of keyframes through shot boundary analysis. From these keyframes, we use state-of- the-art scene detectors and descriptors to extract features, which are then used for shot categorization via unsuper- vised learning. This allows us to represent trailers using a bag-of-visual-words (bovw) model with shot classes as vo- cabularies. We approach the genre classification task by mapping bovw temporally structured trailer features to four high-level movie genres: action, comedy, drama or horror films. We have conducted experiments on 1239 annotated trailers. Our experimental results demonstrate that exploit- ing scene structures improves film genre classification com- pared to using only low-level visual features.

In this paper, we present an example-based system for terrain synthesis. In our approach, patches from a sample terrain (represented by a height field) are used to generate a new terrain. The synthesis is guided by a user-sketched feature map that specifies where terrain features occur in the resulting synthetic terrain. Our system emphasizes large-scale curvilinear features (ridges and valleys) because such features are the dominant visual elements in most terrains. Both the example height field and user's sketch map are analyzed using a technique from the field of geomorphology. The system finds patches from the example data that match the features found in the user's sketch. Patches are joined together using graph cuts and Poisson editing. The order in which patches are placed in the synthesized terrain is determined by breadth-first traversal of a feature tree and this generates improved results over standard raster-scan placement orders. Our technique supports user-controlled terrain synthesis in a wide variety of styles, based upon the visual richness of real-world terrain data.