Spatial understanding is a crucial capability that enables robots to perceive their surroundings, reason about their environment, and interact with it meaningfully. In modern robotics, these capabilities are increasingly provided by vision-language models. However, these models face significant challenges in spatial reasoning tasks, as their training data are based on general-purpose image datasets that often lack sophisticated spatial understanding. For example, datasets frequently do not capture reference frame comprehension, yet effective spatial reasoning requires understanding whether to reason from ego-, world-, or object-centric perspectives. To address this issue, we introduce RoboSpatial, a large-scale dataset for spatial understanding in robotics. It consists of real indoor and tabletop scenes, captured as 3D scans and egocentric images, and annotated with rich spatial information relevant to robotics. The dataset includes 1M images, 5k 3D scans, and 3M annotated spatial relationships, and the pairing of 2D egocentric images with 3D scans makes it both 2D- and 3D- ready. Our experiments show that models trained with RoboSpatial outperform baselines on downstream tasks such as spatial affordance prediction, spatial relationship prediction, and robot manipulation.
The recent development in multimodal learning has greatly advanced the research in 3D scene understanding in various real-world tasks such as embodied AI. However, most existing work shares two typical constraints: 1) they are short of reasoning ability for interaction and interpretation of human intension and 2) they focus on scenarios with single-category objects only which leads to over-simplified textual descriptions due to the negligence of multi-object scenarios and spatial relations among objects. We bridge the research gaps by proposing a 3D reasoning segmentation task for multiple objects in scenes. The task allows producing 3D segmentation masks and detailed textual explanations as enriched by 3D spatial relations among objects. To this end, we create ReasonSeg3D, a large-scale and high-quality benchmark that integrates 3D spatial relations with generated question-answer pairs and 3D segmentation masks. In addition, we design MORE3D, a simple yet effective method that enables multi-object 3D reasoning segmentation with user questions and textual outputs. Extensive experiments show that MORE3D excels in reasoning and segmenting complex multi-object 3D scenes, and the created ReasonSeg3D offers a valuable platform for future exploration of 3D reasoning segmentation. The dataset and code will be released.
Self-supervised learning has made substantial strides in image processing, while visual pre-training for autonomous driving is still in its infancy. Existing methods often focus on learning geometric scene information while neglecting texture or treating both aspects separately, hindering comprehensive scene understanding. In this context, we are excited to introduce GaussianPretrain, a novel pre-training paradigm that achieves a holistic understanding of the scene by uniformly integrating geometric and texture representations. Conceptualizing 3D Gaussian anchors as volumetric LiDAR points, our method learns a deepened understanding of scenes to enhance pre-training performance with detailed spatial structure and texture, achieving that 40.6% faster than NeRF-based method UniPAD with 70% GPU memory only. We demonstrate the effectiveness of GaussianPretrain across multiple 3D perception tasks, showing significant performance improvements, such as a 7.05% increase in NDS for 3D object detection, boosts mAP by 1.9% in HD map construction and 0.8% improvement on Occupancy prediction. These significant gains highlight GaussianPretrain's theoretical innovation and strong practical potential, promoting visual pre-training development for autonomous driving. Source code will be available at https://github.com/Public-BOTs/GaussianPretrain
Satellite imagery has dramatically revolutionized the field of geography by giving academics, scientists, and policymakers unprecedented global access to spatial data. Manual methods typically require significant time and effort to detect the generic land structure in satellite images. This study can produce a set of applications such as urban planning and development, environmental monitoring, disaster management, etc. Therefore, the research presents a methodology to minimize human labor, reducing the expenses and duration needed to identify the land structure. This article developed a deep learning-based approach to automate the process of classifying geographical land structures. We used a satellite image dataset acquired from MLRSNet. The study compared the performance of three architectures, namely CNN, ResNet-50, and Inception-v3. We used three optimizers with any model: Adam, SGD, and RMSProp. We conduct the training process for a fixed number of epochs, specifically 100 epochs, with a batch size of 64. The ResNet-50 achieved an accuracy of 76.5% with the ADAM optimizer, the Inception-v3 with RMSProp achieved an accuracy of 93.8%, and the proposed approach, CNN with RMSProp optimizer, achieved the highest level of performance and an accuracy of 94.8%. Moreover, a thorough examination of the CNN model demonstrated its exceptional accuracy, recall, and F1 scores for all categories, confirming its resilience and dependability in precisely detecting various terrain formations. The results highlight the potential of deep learning models in scene understanding, as well as their significance in efficiently identifying and categorizing land structures from satellite imagery.
Reliable deep learning models require not only accurate predictions but also well-calibrated confidence estimates to ensure dependable uncertainty estimation. This is crucial in safety-critical applications like autonomous driving, which depend on rapid and precise semantic segmentation of LiDAR point clouds for real-time 3D scene understanding. In this work, we introduce a sampling-free approach for estimating well-calibrated confidence values for classification tasks, achieving alignment with true classification accuracy and significantly reducing inference time compared to sampling-based methods. Our evaluation using the Adaptive Calibration Error (ACE) metric for LiDAR semantic segmentation shows that our approach maintains well-calibrated confidence values while achieving increased processing speed compared to a sampling baseline. Additionally, reliability diagrams reveal that our method produces underconfidence rather than overconfident predictions, an advantage for safety-critical applications. Our sampling-free approach offers well-calibrated and time-efficient predictions for LiDAR scene semantic segmentation.
This study introduces a novel approach for traffic control systems by using Large Language Models (LLMs) as traffic controllers. The study utilizes their logical reasoning, scene understanding, and decision-making capabilities to optimize throughput and provide feedback based on traffic conditions in real-time. LLMs centralize traditionally disconnected traffic control processes and can integrate traffic data from diverse sources to provide context-aware decisions. LLMs can also deliver tailored outputs using various means such as wireless signals and visuals to drivers, infrastructures, and autonomous vehicles. To evaluate LLMs ability as traffic controllers, this study proposed a four-stage methodology. The methodology includes data creation and environment initialization, prompt engineering, conflict identification, and fine-tuning. We simulated multi-lane four-leg intersection scenarios and generates detailed datasets to enable conflict detection using LLMs and Python simulation as a ground truth. We used chain-of-thought prompts to lead LLMs in understanding the context, detecting conflicts, resolving them using traffic rules, and delivering context-sensitive traffic management solutions. We evaluated the prformance GPT-mini, Gemini, and Llama as traffic controllers. Results showed that the fine-tuned GPT-mini achieved 83% accuracy and an F1-score of 0.84. GPT-mini model exhibited a promising performance in generating actionable traffic management insights, with high ROUGE-L scores across conflict identification of 0.95, decision-making of 0.91, priority assignment of 0.94, and waiting time optimization of 0.92. We demonstrated that LLMs can offer precise recommendations to drivers in real-time including yielding, slowing, or stopping based on vehicle dynamics.
Scene graphs have proven to be highly effective for various scene understanding tasks due to their compact and explicit representation of relational information. However, current methods often overlook the critical importance of preserving symmetry when generating scene graphs from 3D point clouds, which can lead to reduced accuracy and robustness, particularly when dealing with noisy, multi-view data. This work, to the best of our knowledge, presents the first implementation of an Equivariant Scene Graph Neural Network (ESGNN) to generate semantic scene graphs from 3D point clouds, specifically for enhanced scene understanding. Furthermore, a significant limitation of prior methods is the absence of temporal modeling to capture time-dependent relationships among dynamically evolving entities within a scene. To address this gap, we introduce a novel temporal layer that leverages the symmetry-preserving properties of ESGNN to fuse scene graphs across multiple sequences into a unified global representation by an approximate graph-matching algorithm. Our combined architecture, termed the Temporal Equivariant Scene Graph Neural Network (TESGNN), not only surpasses existing state-of-the-art methods in scene estimation accuracy but also achieves faster convergence. Importantly, TESGNN is computationally efficient and straightforward to implement using existing frameworks, making it well-suited for real-time applications in robotics and computer vision. This approach paves the way for more robust and scalable solutions to complex multi-view scene understanding challenges. Our source code is publicly available at: https://github.com/HySonLab/TESGraph
Underwater surveys provide long-term data for informing management strategies, monitoring coral reef health, and estimating blue carbon stocks. Advances in broad-scale survey methods, such as robotic underwater vehicles, have increased the range of marine surveys but generate large volumes of imagery requiring analysis. Computer vision methods such as semantic segmentation aid automated image analysis, but typically rely on fully supervised training with extensive labelled data. While ground truth label masks for tasks like street scene segmentation can be quickly and affordably generated by non-experts through crowdsourcing services like Amazon Mechanical Turk, ecology presents greater challenges. The complexity of underwater images, coupled with the specialist expertise needed to accurately identify species at the pixel level, makes this process costly, time-consuming, and heavily dependent on domain experts. In recent years, some works have performed automated analysis of underwater imagery, and a smaller number of studies have focused on weakly supervised approaches which aim to reduce the expert-provided labelled data required. This survey focuses on approaches which reduce dependency on human expert input, while reviewing the prior and related approaches to position these works in the wider field of underwater perception. Further, we offer an overview of coastal ecosystems and the challenges of underwater imagery. We provide background on weakly and self-supervised deep learning and integrate these elements into a taxonomy that centres on the intersection of underwater monitoring, computer vision, and deep learning, while motivating approaches for weakly supervised deep learning with reduced dependency on domain expert data annotations. Lastly, the survey examines available datasets and platforms, and identifies gaps, barriers, and opportunities for automating underwater surveys.
Comprehensively understanding surgical scenes in Surgical Visual Question Answering (Surgical VQA) requires reasoning over multiple objects. Previous approaches address this task using cross-modal fusion strategies to enhance reasoning ability. However, these methods often struggle with limited scene understanding and question comprehension, and some rely on external resources (e.g., pre-extracted object features), which can introduce errors and generalize poorly across diverse surgical environments. To address these challenges, we propose SCAN, a simple yet effective memory-augmented framework that leverages Multimodal LLMs to improve surgical context comprehension via Self-Contained Inquiry. SCAN operates autonomously, generating two types of memory for context augmentation: Direct Memory (DM), which provides multiple candidates (or hints) to the final answer, and Indirect Memory (IM), which consists of self-contained question-hint pairs to capture broader scene context. DM directly assists in answering the question, while IM enhances understanding of the surgical scene beyond the immediate query. Reasoning over these object-aware memories enables the model to accurately interpret images and respond to questions. Extensive experiments on three publicly available Surgical VQA datasets demonstrate that SCAN achieves state-of-the-art performance, offering improved accuracy and robustness across various surgical scenarios.
Monocular geometric scene understanding combines panoptic segmentation and self-supervised depth estimation, focusing on real-time application in autonomous vehicles. We introduce MGNiceNet, a unified approach that uses a linked kernel formulation for panoptic segmentation and self-supervised depth estimation. MGNiceNet is based on the state-of-the-art real-time panoptic segmentation method RT-K-Net and extends the architecture to cover both panoptic segmentation and self-supervised monocular depth estimation. To this end, we introduce a tightly coupled self-supervised depth estimation predictor that explicitly uses information from the panoptic path for depth prediction. Furthermore, we introduce a panoptic-guided motion masking method to improve depth estimation without relying on video panoptic segmentation annotations. We evaluate our method on two popular autonomous driving datasets, Cityscapes and KITTI. Our model shows state-of-the-art results compared to other real-time methods and closes the gap to computationally more demanding methods. Source code and trained models are available at https://github.com/markusschoen/MGNiceNet.
Depth estimation is an essential task toward full scene understanding since it allows the projection of rich semantic information captured by cameras into 3D space. While the field has gained much attention recently, datasets for depth estimation lack scene diversity or sensor modalities. This work presents the ADUULM-360 dataset, a novel multi-modal dataset for depth estimation. The ADUULM-360 dataset covers all established autonomous driving sensor modalities, cameras, lidars, and radars. It covers a frontal-facing stereo setup, six surround cameras covering the full 360-degree, two high-resolution long-range lidar sensors, and five long-range radar sensors. It is also the first depth estimation dataset that contains diverse scenes in good and adverse weather conditions. We conduct extensive experiments using state-of-the-art self-supervised depth estimation methods under different training tasks, such as monocular training, stereo training, and full surround training. Discussing these results, we demonstrate common limitations of state-of-the-art methods, especially in adverse weather conditions, which hopefully will inspire future research in this area. Our dataset, development kit, and trained baselines are available at https://github.com/uulm-mrm/aduulm_360_dataset.
Recovering metric depth from a single image remains a fundamental challenge in computer vision, requiring both scene understanding and accurate scaling. While deep learning has advanced monocular depth estimation, current models often struggle with unfamiliar scenes and layouts, particularly in zero-shot scenarios and when predicting scale-ergodic metric depth. We present MetricGold, a novel approach that harnesses generative diffusion model's rich priors to improve metric depth estimation. Building upon recent advances in MariGold, DDVM and Depth Anything V2 respectively, our method combines latent diffusion, log-scaled metric depth representation, and synthetic data training. MetricGold achieves efficient training on a single RTX 3090 within two days using photo-realistic synthetic data from HyperSIM, VirtualKitti, and TartanAir. Our experiments demonstrate robust generalization across diverse datasets, producing sharper and higher quality metric depth estimates compared to existing approaches.
Remarkable progress has been achieved in image generation with the introduction of generative models. However, precisely controlling the content in generated images remains a challenging task due to their fundamental training objective. This paper addresses this challenge by proposing a novel image generation framework explicitly designed to incorporate desired content in output images. The framework utilizes advanced encoding techniques, integrating subnetworks called content fusion and frequency encoding modules. The frequency encoding module first captures features and structures of reference images by exclusively focusing on selected frequency components. Subsequently, the content fusion module generates a content-guiding vector that encapsulates desired content features. During the image generation process, content-guiding vectors from real images are fused with projected noise vectors. This ensures the production of generated images that not only maintain consistent content from guiding images but also exhibit diverse stylistic variations. To validate the effectiveness of the proposed framework in preserving content attributes, extensive experiments are conducted on widely used benchmark datasets, including Flickr-Faces-High Quality, Animal Faces High Quality, and Large-scale Scene Understanding datasets.
OpenStreetMap (OSM), a rich and versatile source of volunteered geographic information (VGI), facilitates human self-localization and scene understanding by integrating nearby visual observations with vectorized map data. However, the disparity in modalities and perspectives poses a major challenge for effectively matching camera imagery with compact map representations, thereby limiting the full potential of VGI data in real-world localization applications. Inspired by the fact that the human brain relies on the fusion of geometric and semantic understanding for spatial localization tasks, we propose the OSMLoc in this paper. OSMLoc is a brain-inspired visual localization approach based on first-person-view images against the OSM maps. It integrates semantic and geometric guidance to significantly improve accuracy, robustness, and generalization capability. First, we equip the OSMLoc with the visual foundational model to extract powerful image features. Second, a geometry-guided depth distribution adapter is proposed to bridge the monocular depth estimation and camera-to-BEV transform. Thirdly, the semantic embeddings from the OSM data are utilized as auxiliary guidance for image-to-OSM feature matching. To validate the proposed OSMLoc, we collect a worldwide cross-area and cross-condition (CC) benchmark for extensive evaluation. Experiments on the MGL dataset, CC validation benchmark, and KITTI dataset have demonstrated the superiority of our method. Code, pre-trained models, CC validation benchmark, and additional results are available at: https://github.com/WHU-USI3DV/OSMLoc.
Incorporating inductive bias by embedding geometric entities (such as rays) as input has proven successful in multi-view learning. However, the methods adopting this technique typically lack equivariance, which is crucial for effective 3D learning. Equivariance serves as a valuable inductive prior, aiding in the generation of robust multi-view features for 3D scene understanding. In this paper, we explore the application of equivariant multi-view learning to depth estimation, not only recognizing its significance for computer vision and robotics but also addressing the limitations of previous research. Most prior studies have either overlooked equivariance in this setting or achieved only approximate equivariance through data augmentation, which often leads to inconsistencies across different reference frames. To address this issue, we propose to embed $SE(3)$ equivariance into the Perceiver IO architecture. We employ Spherical Harmonics for positional encoding to ensure 3D rotation equivariance, and develop a specialized equivariant encoder and decoder within the Perceiver IO architecture. To validate our model, we applied it to the task of stereo depth estimation, achieving state of the art results on real-world datasets without explicit geometric constraints or extensive data augmentation.
The growing demand for robust scene understanding in mobile robotics and autonomous driving has highlighted the importance of integrating multiple sensing modalities. By combining data from diverse sensors like cameras and LIDARs, fusion techniques can overcome the limitations of individual sensors, enabling a more complete and accurate perception of the environment. We introduce a novel approach to multi-modal sensor fusion, focusing on developing a graph-based state representation that supports critical decision-making processes in autonomous driving. We present a Sensor-Agnostic Graph-Aware Kalman Filter [3], the first online state estimation technique designed to fuse multi-modal graphs derived from noisy multi-sensor data. The estimated graph-based state representations serve as a foundation for advanced applications like Multi-Object Tracking (MOT), offering a comprehensive framework for enhancing the situational awareness and safety of autonomous systems. We validate the effectiveness of our proposed framework through extensive experiments conducted on both synthetic and real-world driving datasets (nuScenes). Our results showcase an improvement in MOTA and a reduction in estimated position errors (MOTP) and identity switches (IDS) for tracked objects using the SAGA-KF. Furthermore, we highlight the capability of such a framework to develop methods that can leverage heterogeneous information (like semantic objects and geometric structures) from various sensing modalities, enabling a more holistic approach to scene understanding and enhancing the safety and effectiveness of autonomous systems.
In this paper, we present a novel algorithm for probabilistically updating and rasterizing semantic maps within 3D Gaussian Splatting (3D-GS). Although previous methods have introduced algorithms which learn to rasterize features in 3D-GS for enhanced scene understanding, 3D-GS can fail without warning which presents a challenge for safety-critical robotic applications. To address this gap, we propose a method which advances the literature of continuous semantic mapping from voxels to ellipsoids, combining the precise structure of 3D-GS with the ability to quantify uncertainty of probabilistic robotic maps. Given a set of images, our algorithm performs a probabilistic semantic update directly on the 3D ellipsoids to obtain an expectation and variance through the use of conjugate priors. We also propose a probabilistic rasterization which returns per-pixel segmentation predictions with quantifiable uncertainty. We compare our method with similar probabilistic voxel-based methods to verify our extension to 3D ellipsoids, and perform ablation studies on uncertainty quantification and temporal smoothing.
Geometric estimation is required for scene understanding and analysis in panoramic 360{\deg} images. Current methods usually predict a single feature, such as depth or surface normal. These methods can lack robustness, especially when dealing with intricate textures or complex object surfaces. We introduce a novel multi-task learning (MTL) network that simultaneously estimates depth and surface normals from 360{\deg} images. Our first innovation is our MTL architecture, which enhances predictions for both tasks by integrating geometric information from depth and surface normal estimation, enabling a deeper understanding of 3D scene structure. Another innovation is our fusion module, which bridges the two tasks, allowing the network to learn shared representations that improve accuracy and robustness. Experimental results demonstrate that our MTL architecture significantly outperforms state-of-the-art methods in both depth and surface normal estimation, showing superior performance in complex and diverse scenes. Our model's effectiveness and generalizability, particularly in handling intricate surface textures, establish it as a new benchmark in 360{\deg} image geometric estimation. The code and model are available at \url{https://github.com/huangkun101230/360MTLGeometricEstimation}.
3D LiDAR point cloud data is crucial for scene perception in computer vision, robotics, and autonomous driving. Geometric and semantic scene understanding, involving 3D point clouds, is essential for advancing autonomous driving technologies. However, significant challenges remain, particularly in improving the overall accuracy (e.g., segmentation accuracy, depth estimation accuracy, etc.) and efficiency of these systems. To address the challenge in terms of accuracy related to LiDAR-based tasks, we present DurLAR, the first high-fidelity 128-channel 3D LiDAR dataset featuring panoramic ambient (near infrared) and reflectivity imagery. To improve efficiency in 3D segmentation while ensuring the accuracy, we propose a novel pipeline that employs a smaller architecture, requiring fewer ground-truth annotations while achieving superior segmentation accuracy compared to contemporary approaches. To improve the segmentation accuracy, we introduce Range-Aware Pointwise Distance Distribution (RAPiD) features and the associated RAPiD-Seg architecture. All contributions have been accepted by peer-reviewed conferences, underscoring the advancements in both accuracy and efficiency in 3D LiDAR applications for autonomous driving. Full abstract: https://etheses.dur.ac.uk/15738/.
Scene understanding is a fundamental capability needed in many domains, ranging from question-answering to robotics. Unlike recent end-to-end approaches that must explicitly learn varying compositions of the same scene, our method reasons over their constituent objects and analyzes their arrangement to infer a scene's meaning. We propose a novel approach that reasons over a scene's scene- and knowledge-graph, capturing spatial information while being able to utilize general domain knowledge in a joint graph search. Empirically, we demonstrate the feasibility of our method on the ADE20K dataset and compare it to current scene understanding approaches.