3D occupancy prediction (3DOcc) is a rapidly rising and challenging perception task in the field of autonomous driving. Existing 3D occupancy networks (OccNets) are both computationally heavy and label-hungry. In terms of model complexity, OccNets are commonly composed of heavy Conv3D modules or transformers at the voxel level. Moreover, OccNets are supervised with expensive large-scale dense voxel labels. Model and label inefficiencies, caused by excessive network parameters and label annotation requirements, severely hinder the onboard deployment of OccNets. This paper proposes an EFFicient Occupancy learning framework, EFFOcc, that targets minimal network complexity and label requirements while achieving state-of-the-art accuracy. We first propose an efficient fusion-based OccNet that only uses simple 2D operators and improves accuracy to the state-of-the-art on three large-scale benchmarks: Occ3D-nuScenes, Occ3D-Waymo, and OpenOccupancy-nuScenes. On the Occ3D-nuScenes benchmark, the fusion-based model with ResNet-18 as the image backbone has 21.35M parameters and achieves 51.49 in terms of mean Intersection over Union (mIoU). Furthermore, we propose a multi-stage occupancy-oriented distillation to efficiently transfer knowledge to vision-only OccNet. Extensive experiments on occupancy benchmarks show state-of-the-art precision for both fusion-based and vision-based OccNets. For the demonstration of learning with limited labels, we achieve 94.38% of the performance (mIoU = 28.38) of a 100% labeled vision OccNet (mIoU = 30.07) using the same OccNet trained with only 40% labeled sequences and distillation from the fusion-based OccNet. Code is available at https://github.com/synsin0/EFFOcc.

World models are critical for autonomous driving to simulate environmental dynamics and generate synthetic data. Existing methods struggle to disentangle ego-vehicle motion (perspective shifts) from scene evolvement (agent interactions), leading to suboptimal predictions. Instead, we propose to separate environmental changes from ego-motion by leveraging the scene-centric coordinate systems. In this paper, we introduce COME: a framework that integrates scene-centric forecasting Control into the Occupancy world ModEl. Specifically, COME first generates ego-irrelevant, spatially consistent future features through a scene-centric prediction branch, which are then converted into scene condition using a tailored ControlNet. These condition features are subsequently injected into the occupancy world model, enabling more accurate and controllable future occupancy predictions. Experimental results on the nuScenes-Occ3D dataset show that COME achieves consistent and significant improvements over state-of-the-art (SOTA) methods across diverse configurations, including different input sources (ground-truth, camera-based, fusion-based occupancy) and prediction horizons (3s and 8s). For example, under the same settings, COME achieves 26.3% better mIoU metric than DOME and 23.7% better mIoU metric than UniScene. These results highlight the efficacy of disentangled representation learning in enhancing spatio-temporal prediction fidelity for world models. Code is available at https://github.com/synsin0/COME.

The grid-centric perception is a crucial field for mobile robot perception and navigation. Nonetheless, the grid-centric perception is less prevalent than object-centric perception as autonomous vehicles need to accurately perceive highly dynamic, large-scale traffic scenarios, and the complexity and computational costs of grid-centric perception are high. In recent years, the rapid development of deep learning techniques and hardware provides fresh insights into the evolution of grid-centric perception. The fundamental difference between grid-centric and object-centric pipeline lies in that grid-centric perception follows a geometry-first paradigm which is more robust to the open-world driving scenarios with endless long-tailed semantically unknown obstacles. Recent research demonstrates the great advantages of grid-centric perception, such as comprehensive fine-grained environmental representation, greater robustness to occlusion and irregular-shaped objects, better ground estimation, and safer planning policies. There is also a growing trend that the capacity of occupancy networks is greatly expanded to 4-D scene perception and prediction, and the latest techniques are highly related to new research topics, such as 4-D occupancy forecasting, generative artificial intelligence (GenAI), and world models in the field of autonomous driving. Given the lack of current surveys for this rapidly expanding field, we present a hierarchically structured review of grid-centric perception for autonomous vehicles. We organize previous and current knowledge of occupancy grid techniques along the main vein from 2-D bird-eye view (BEV) grids to 3-D occupancy to 4-D occupancy forecasting. We additionally summarize label-efficient occupancy learning and the role of grid-centric perception in driving systems. Finally, we present a summary of the current research trend and provide future outlooks.

Vision-centric occupancy networks, which represent the surrounding environment with uniform voxels with semantics, have become a new trend for safe driving of camera-only autonomous driving perception systems, as they are able to detect obstacles regardless of their shape and occlusion. Modern occupancy networks mainly focus on reconstructing visible voxels from object surfaces with voxel-wise semantic prediction. Usually, they suffer from inconsistent predictions of one object and mixed predictions for adjacent objects. These confusions may harm the safety of downstream planning modules. To this end, we investigate panoptic segmentation on 3D voxel scenarios and propose an instance-aware occupancy network, PanoSSC. We predict foreground objects and backgrounds separately and merge both in post-processing. For foreground instance grouping, we propose a novel 3D instance mask decoder that can efficiently extract individual objects. we unify geometric reconstruction, 3D semantic segmentation, and 3D instance segmentation into PanoSSC framework and propose new metrics for evaluating panoptic voxels. Extensive experiments show that our method achieves competitive results on SemanticKITTI semantic scene completion benchmark.

Predicting the future occupancy states of the surrounding environment is a vital task for autonomous driving. However, current best-performing single-modality methods or multi-modality fusion perception methods are only able to predict uniform snapshots of future occupancy states and require strictly synchronized sensory data for sensor fusion. We propose a novel framework, StreamingFlow, to lift these strong limitations. StreamingFlow is a novel BEV occupancy predictor that ingests asynchronous multi-sensor data streams for fusion and performs streaming fore-casting of the future occupancy map at any future times-tamps. By integrating neural ordinary differential equations (N-ODE) into recurrent neural networks, StreamingFlow learns derivatives of BEV features over temporal horizons, updates the implicit sensor's BEV features as part of the fusion process, and propagates BEV states to the desired future time point. It shows good zero-shot generalization ability of prediction, reflected in the interpolation of the ob-served prediction time horizon and the reasonable inference of the unseen farther future period. Extensive experiments on two large-scale datasets, nuScenes [2] and Lyft L5 [14], demonstrate that StreamingFlow significantly outperforms previous vision-based, LiDAR-based methods, and shows superior performance compared to state-of-the-art fusion-based methods.

If a picture is worth a thousand words, then images should be utilized together with other available data modalities when predicting the virality of online posts, such as tweets. In this paper, we re-visit the tweet popularity prediction problem by considering all data modalities: tweet language semantics, embedded images, author' social relationships, and the diffusion process of tweets. To model the content of tweets, we propose a joint-embedding neural network that combines visual, textual, and social cues together. Such content features can be either used for prediction directly, or for pre-conditioning a 'dynamics RNN', which models the message propagation process. A novel Poisson regression loss is optimized to train the network. We demonstrate that content based features can be used to improve upon social features and dynamics features via our joint-embedding regression model. Our model outperforms the state-of-the-art on multiple large-scale real-world datasets collected from Twitter.

Most recent garment capturing techniques rely on acquiring multiple views of clothing, which may not always be readily available, especially in the case of pre-existing photographs from the web. As an alternative, we propose a method that is able to compute a 3D model of a human body and its outfit from a single photograph with little human interaction. Our algorithm is not only able to capture the global shape and overall geometry of the clothing, it can also extract the physical properties (i.e., material parameters needed for simulation) of cloth. Unlike previous methods using full 3D information (i.e., depth, multi-view images, or sampled 3D geometry), our approach achieves garment recovery from a single-view image by using physical, statistical, and geometric priors and a combination of parameter estimation, semantic parsing, shape/pose recovery, and physics-based cloth simulation. We demonstrate the effectiveness of our algorithm by re-purposing the reconstructed garments for virtual try-on and garment transfer applications and for cloth animation on digital characters.

In this paper, we propose a fast and accurate approach for 3D reconstructions from satellite images. Compared with traditional images, satellite imagery features enormous pixel count, inaccurate camera calibration, and low ground sampling rate, all of which makes multi-view stereo for satellite images more challenging. Our approach first computes sparse but reliable 2D feature matches between image pairs. Such feature matches are used to compensate the extrinsic calibration errors. Preliminary dense correspondences are obtained via edge-aware interpolation of sparse feature matches. We rely on fast bilateral smoothing to refine such initial dense matches, which greatly improves the computational efficiency of our method. The smoothed dense correspondences and the refined camera model are then used to obtain dense 3D point clouds via triangulation. Our proposed method outperforms state-of-the-art baseline methods in both efficiency and accuracy on real-world datasets.

Satellite images have broad coverage, thus are ideal for large-scale urban reconstruction tasks. However, their low ground sampling resolution posed great challenges in using traditional volumetric or stereo methods to perform 3D reconstructions. In this paper, we propose a novel deep learning based approach to perform single-view parametric reconstructions from satellite imagery. By parametrizing buildings as 3D cuboids, our method extends object detection systems to simultaneously localize buildings and directly fit parametric models for each identified building. We utilize geo-registered GIS vector maps and Lidar data as supervision to train the network. Especially, we deconvolve the feature maps and combine convolutional feature maps at different stages of the network to deal with the heavily cluttered but small in size building instances from satellite imagery. We further enforce physical constraints that building cannot overlap by predicting building boundaries using a separate fully convolutional network. We demonstrate the effectiveness of our proposed methods on real-world data.

We propose an efficient algorithm to jointly estimate geometry and semantics for a given geographical region observed by multiple satellite images. Our joint estimation leverages an efficient PatchMatch inference framework defined over lattice discretization of the environment. Our cost function relies on the local planarity assumption to model scene geometry and neural network classification to determine semantic (e.g. land use) labels for geometric structures. By utilizing the commonly available direct (i.e. space to image) rational polynomial coefficients (RPC) satellite camera models, our approach effectively circumvents the need for estimating or refining inverse RPC models. Experiments illustrate both the computational efficiency and high quality scene geometry estimates attained by our approach for satellite imagery. To further illustrate the generality of our representation and inference framework, experiments on standard benchmarks for ground-level imagery are also included.

We propose two novel minimal solvers which advance the state of the art in satellite imagery processing. Our methods are efficient and do not rely on the prior existence of complex inverse mapping functions to correlate 2D image coordinates and 3D terrain. Our first solver improves on the stereo correspondence problem for satellite imagery, in that we provide an exact image-to-object space mapping (where prior methods were inaccurate). Our second solver provides a novel mechanism for 3D point triangulation, which has improved robustness and accuracy over prior techniques. Given the usefulness and ubiquity of satellite imagery, our proposed methods allow for improved results in a variety of existing and future applications.

We develop a sequential optimal sampling framework for stereo disparity estimation by adapting the Sequential Probability Ratio Test (SPRT) model. We operate over local image neighborhoods by iteratively estimating single pixel disparity values until sufficient evidence has been gathered to either validate or contradict the current hypothesis regarding local scene structure. The output of our sampling is a set of sampled pixel positions along with a robust and compact estimate of the set of disparities contained within a given region. We further propose an efficient plane propagation mechanism that leverages the pre-computed sampling positions and the local structure model described by the reduced local disparity set. Our sampling framework is a general pre-processing mechanism aimed at reducing computational complexity of disparity search algorithms by ascertaining a reduced set of disparity hypotheses for each pixel. Experiments demonstrate the effectiveness of the proposed approach when compared to state of the art methods.

We consider the problem of estimating the relative depth of a scene from a monocular image. The dark channel prior, used as a statistical observation of haze free images, has been previously leveraged for haze removal and relative depth estimation tasks. However, as a local measure, it fails to account for higher order semantic relationship among scene elements. We propose a dual channel prior used for identifying pixels that are unlikely to comply with the dark channel assumption, leading to erroneous depth estimates. We further leverage semantic segmentation information and patch match label propagation to enforce semantically consistent geometric priors. Experiments illustrate the quantitative and qualitative advantages of our approach when compared to state of the art methods.