There is increased interest in using generative AI to create 3D spaces for Virtual Reality (VR) applications. However, today's models produce artificial environments, falling short of supporting collaborative tasks that benefit from incorporating the user's physical context. To generate environments that support VR telepresence, we introduce SpaceBlender, a novel pipeline that utilizes generative AI techniques to blend users' physical surroundings into unified virtual spaces. This pipeline transforms user-provided 2D images into context-rich 3D environments through an iterative process consisting of depth estimation, mesh alignment, and diffusion-based space completion guided by geometric priors and adaptive text prompts. In a preliminary within-subjects study, where 20 participants performed a collaborative VR affinity diagramming task in pairs, we compared SpaceBlender with a generic virtual environment and a state-of-the-art scene generation framework, evaluating its ability to create virtual spaces suitable for collaboration. Participants appreciated the enhanced familiarity and context provided by SpaceBlender but also noted complexities in the generative environments that could detract from task focus. Drawing on participant feedback, we propose directions for improving the pipeline and discuss the value and design of blended spaces for different scenarios.
We present Uni-DVPS, a unified model for Depth-aware Video Panoptic Segmentation (DVPS) that jointly tackles distinct vision tasks, i.e., video panoptic segmentation, monocular depth estimation, and object tracking. In contrast to the prior works that adopt diverged decoder networks tailored for each task, we propose an architecture with a unified Transformer decoder network. We design a single Transformer decoder network for multi-task learning to increase shared operations to facilitate the synergies between tasks and exhibit high efficiency. We also observe that our unified query learns instance-aware representation guided by multi-task supervision, which encourages query-based tracking and obviates the need for training extra tracking module. We validate our architectural design choices with experiments on Cityscapes-DVPS and SemKITTI-DVPS datasets. The performances of all tasks are jointly improved, and we achieve state-of-the-art results on DVPQ metric for both datasets.
Vision-based 3D occupancy prediction is significantly challenged by the inherent limitations of monocular vision in depth estimation. This paper introduces CVT-Occ, a novel approach that leverages temporal fusion through the geometric correspondence of voxels over time to improve the accuracy of 3D occupancy predictions. By sampling points along the line of sight of each voxel and integrating the features of these points from historical frames, we construct a cost volume feature map that refines current volume features for improved prediction outcomes. Our method takes advantage of parallax cues from historical observations and employs a data-driven approach to learn the cost volume. We validate the effectiveness of CVT-Occ through rigorous experiments on the Occ3D-Waymo dataset, where it outperforms state-of-the-art methods in 3D occupancy prediction with minimal additional computational cost. The code is released at \url{https://github.com/Tsinghua-MARS-Lab/CVT-Occ}.
Depth estimation is a fundamental task in 3D geometry. While stereo depth estimation can be achieved through triangulation methods, it is not as straightforward for monocular methods, which require the integration of global and local information. The Depth from Defocus (DFD) method utilizes camera lens models and parameters to recover depth information from blurred images and has been proven to perform well. However, these methods rely on All-In-Focus (AIF) images for depth estimation, which is nearly impossible to obtain in real-world applications. To address this issue, we propose a self-supervised framework based on 3D Gaussian splatting and Siamese networks. By learning the blur levels at different focal distances of the same scene in the focal stack, the framework predicts the defocus map and Circle of Confusion (CoC) from a single defocused image, using the defocus map as input to DepthNet for monocular depth estimation. The 3D Gaussian splatting model renders defocused images using the predicted CoC, and the differences between these and the real defocused images provide additional supervision signals for the Siamese Defocus self-supervised network. This framework has been validated on both artificially synthesized and real blurred datasets. Subsequent quantitative and visualization experiments demonstrate that our proposed framework is highly effective as a DFD method.
Forecasting the semantics and 3D structure of scenes is essential for robots to navigate and plan actions safely. Recent methods have explored semantic and panoptic scene forecasting; however, they do not consider the geometry of the scene. In this work, we propose the panoptic-depth forecasting task for jointly predicting the panoptic segmentation and depth maps of unobserved future frames, from monocular camera images. To facilitate this work, we extend the popular KITTI-360 and Cityscapes benchmarks by computing depth maps from LiDAR point clouds and leveraging sequential labeled data. We also introduce a suitable evaluation metric that quantifies both the panoptic quality and depth estimation accuracy of forecasts in a coherent manner. Furthermore, we present two baselines and propose the novel PDcast architecture that learns rich spatio-temporal representations by incorporating a transformer-based encoder, a forecasting module, and task-specific decoders to predict future panoptic-depth outputs. Extensive evaluations demonstrate the effectiveness of PDcast across two datasets and three forecasting tasks, consistently addressing the primary challenges. We make the code publicly available at https://pdcast.cs.uni-freiburg.de.
Recent work showed that large diffusion models can be reused as highly precise monocular depth estimators by casting depth estimation as an image-conditional image generation task. While the proposed model achieved state-of-the-art results, high computational demands due to multi-step inference limited its use in many scenarios. In this paper, we show that the perceived inefficiency was caused by a flaw in the inference pipeline that has so far gone unnoticed. The fixed model performs comparably to the best previously reported configuration while being more than 200$\times$ faster. To optimize for downstream task performance, we perform end-to-end fine-tuning on top of the single-step model with task-specific losses and get a deterministic model that outperforms all other diffusion-based depth and normal estimation models on common zero-shot benchmarks. We surprisingly find that this fine-tuning protocol also works directly on Stable Diffusion and achieves comparable performance to current state-of-the-art diffusion-based depth and normal estimation models, calling into question some of the conclusions drawn from prior works.
This work addresses the task of zero-shot monocular depth estimation. A recent advance in this field has been the idea of utilising Text-to-Image foundation models, such as Stable Diffusion. Foundation models provide a rich and generic image representation, and therefore, little training data is required to reformulate them as a depth estimation model that predicts highly-detailed depth maps and has good generalisation capabilities. However, the realisation of this idea has so far led to approaches which are, unfortunately, highly inefficient at test-time due to the underlying iterative denoising process. In this work, we propose a different realisation of this idea and present PrimeDepth, a method that is highly efficient at test time while keeping, or even enhancing, the positive aspects of diffusion-based approaches. Our key idea is to extract from Stable Diffusion a rich, but frozen, image representation by running a single denoising step. This representation, we term preimage, is then fed into a refiner network with an architectural inductive bias, before entering the downstream task. We validate experimentally that PrimeDepth is two orders of magnitude faster than the leading diffusion-based method, Marigold, while being more robust for challenging scenarios and quantitatively marginally superior. Thereby, we reduce the gap to the currently leading data-driven approach, Depth Anything, which is still quantitatively superior, but predicts less detailed depth maps and requires 20 times more labelled data. Due to the complementary nature of our approach, even a simple averaging between PrimeDepth and Depth Anything predictions can improve upon both methods and sets a new state-of-the-art in zero-shot monocular depth estimation. In future, data-driven approaches may also benefit from integrating our preimage.
3D reconstruction from a single image is a long-standing problem in computer vision. Learning-based methods address its inherent scale ambiguity by leveraging increasingly large labeled and unlabeled datasets, to produce geometric priors capable of generating accurate predictions across domains. As a result, state of the art approaches show impressive performance in zero-shot relative and metric depth estimation. Recently, diffusion models have exhibited remarkable scalability and generalizable properties in their learned representations. However, because these models repurpose tools originally designed for image generation, they can only operate on dense ground-truth, which is not available for most depth labels, especially in real-world settings. In this paper we present GRIN, an efficient diffusion model designed to ingest sparse unstructured training data. We use image features with 3D geometric positional encodings to condition the diffusion process both globally and locally, generating depth predictions at a pixel-level. With comprehensive experiments across eight indoor and outdoor datasets, we show that GRIN establishes a new state of the art in zero-shot metric monocular depth estimation even when trained from scratch.
Controllable Depth-of-Field (DoF) imaging commonly produces amazing visual effects based on heavy and expensive high-end lenses. However, confronted with the increasing demand for mobile scenarios, it is desirable to achieve a lightweight solution with Minimalist Optical Systems (MOS). This work centers around two major limitations of MOS, i.e., the severe optical aberrations and uncontrollable DoF, for achieving single-lens controllable DoF imaging via computational methods. A Depth-aware Controllable DoF Imaging (DCDI) framework is proposed equipped with All-in-Focus (AiF) aberration correction and monocular depth estimation, where the recovered image and corresponding depth map are utilized to produce imaging results under diverse DoFs of any high-end lens via patch-wise convolution. To address the depth-varying optical degradation, we introduce a Depth-aware Degradation-adaptive Training (DA2T) scheme. At the dataset level, a Depth-aware Aberration MOS (DAMOS) dataset is established based on the simulation of Point Spread Functions (PSFs) under different object distances. Additionally, we design two plug-and-play depth-aware mechanisms to embed depth information into the aberration image recovery for better tackling depth-aware degradation. Furthermore, we propose a storage-efficient Omni-Lens-Field model to represent the 4D PSF library of various lenses. With the predicted depth map, recovered image, and depth-aware PSF map inferred by Omni-Lens-Field, single-lens controllable DoF imaging is achieved. Comprehensive experimental results demonstrate that the proposed framework enhances the recovery performance, and attains impressive single-lens controllable DoF imaging results, providing a seminal baseline for this field. The source code and the established dataset will be publicly available at https://github.com/XiaolongQian/DCDI.
Traditional fish farming practices often lead to inefficient feeding, resulting in environmental issues and reduced productivity. We developed an innovative system combining computer vision and IoT technologies for precise Tilapia feeding. Our solution uses real-time IoT sensors to monitor water quality parameters and computer vision algorithms to analyze fish size and count, determining optimal feed amounts. A mobile app enables remote monitoring and control. We utilized YOLOv8 for keypoint detection to measure Tilapia weight from length, achieving \textbf{94\%} precision on 3,500 annotated images. Pixel-based measurements were converted to centimeters using depth estimation for accurate feeding calculations. Our method, with data collection mirroring inference conditions, significantly improved results. Preliminary estimates suggest this approach could increase production up to 58 times compared to traditional farms. Our models, code, and dataset are open-source~\footnote{The code, dataset, and models are available upon reasonable request.
Nighttime camera-based depth estimation is a highly challenging task, especially for autonomous driving applications, where accurate depth perception is essential for ensuring safe navigation. We aim to improve the reliability of perception systems at night time, where models trained on daytime data often fail in the absence of precise but costly LiDAR sensors. In this work, we introduce Light Enhanced Depth (LED), a novel cost-effective approach that significantly improves depth estimation in low-light environments by harnessing a pattern projected by high definition headlights available in modern vehicles. LED leads to significant performance boosts across multiple depth-estimation architectures (encoder-decoder, Adabins, DepthFormer) both on synthetic and real datasets. Furthermore, increased performances beyond illuminated areas reveal a holistic enhancement in scene understanding. Finally, we release the Nighttime Synthetic Drive Dataset, a new synthetic and photo-realistic nighttime dataset, which comprises 49,990 comprehensively annotated images.
Omnidirectional Depth Estimation has broad application prospects in fields such as robotic navigation and autonomous driving. In this paper, we propose a robotic prototype system and corresponding algorithm designed to validate omnidirectional depth estimation for navigation and obstacle avoidance in real-world scenarios for both robots and vehicles. The proposed HexaMODE system captures 360$^\circ$ depth maps using six surrounding arranged fisheye cameras. We introduce a combined spherical sweeping method and optimize the model architecture for proposed RtHexa-OmniMVS algorithm to achieve real-time omnidirectional depth estimation. To ensure high accuracy, robustness, and generalization in real-world environments, we employ a teacher-student self-training strategy, utilizing large-scale unlabeled real-world data for model training. The proposed algorithm demonstrates high accuracy in various complex real-world scenarios, both indoors and outdoors, achieving an inference speed of 15 fps on edge computing platforms.
High frame rate and accurate depth estimation plays an important role in several tasks crucial to robotics and automotive perception. To date, this can be achieved through ToF and LiDAR devices for indoor and outdoor applications, respectively. However, their applicability is limited by low frame rate, energy consumption, and spatial sparsity. Depth on Demand (DoD) allows for accurate temporal and spatial depth densification achieved by exploiting a high frame rate RGB sensor coupled with a potentially lower frame rate and sparse active depth sensor. Our proposal jointly enables lower energy consumption and denser shape reconstruction, by significantly reducing the streaming requirements on the depth sensor thanks to its three core stages: i) multi-modal encoding, ii) iterative multi-modal integration, and iii) depth decoding. We present extended evidence assessing the effectiveness of DoD on indoor and outdoor video datasets, covering both environment scanning and automotive perception use cases.
Depth estimation is a cornerstone of 3D reconstruction and plays a vital role in minimally invasive endoscopic surgeries. However, most current depth estimation networks rely on traditional convolutional neural networks, which are limited in their ability to capture global information. Foundation models offer a promising avenue for enhancing depth estimation, but those currently available are primarily trained on natural images, leading to suboptimal performance when applied to endoscopic images. In this work, we introduce a novel fine-tuning strategy for the Depth Anything Model and integrate it with an intrinsic-based unsupervised monocular depth estimation framework. Our approach includes a low-rank adaptation technique based on random vectors, which improves the model's adaptability to different scales. Additionally, we propose a residual block built on depthwise separable convolution to compensate for the transformer's limited ability to capture high-frequency details, such as edges and textures. Our experimental results on the SCARED dataset show that our method achieves state-of-the-art performance while minimizing the number of trainable parameters. Applying this method in minimally invasive endoscopic surgery could significantly enhance both the precision and safety of these procedures.
Robust depth perception in visually-degraded environments is crucial for autonomous aerial systems. Thermal imaging cameras, which capture infrared radiation, are robust to visual degradation. However, due to lack of a large-scale dataset, the use of thermal cameras for unmanned aerial system (UAS) depth perception has remained largely unexplored. This paper presents a stereo thermal depth perception dataset for autonomous aerial perception applications. The dataset consists of stereo thermal images, LiDAR, IMU and ground truth depth maps captured in urban and forest settings under diverse conditions like day, night, rain, and smoke. We benchmark representative stereo depth estimation algorithms, offering insights into their performance in degraded conditions. Models trained on our dataset generalize well to unseen smoky conditions, highlighting the robustness of stereo thermal imaging for depth perception. We aim for this work to enhance robotic perception in disaster scenarios, allowing for exploration and operations in previously unreachable areas. The dataset and source code are available at https://firestereo.github.io.
We present a fully constructive analysis of deep ReLU neural networks for classification and function approximation tasks. First, we prove that any dataset with $N$ distinct points in $\mathbb{R}^d$ and $M$ output classes can be exactly classified using a multilayer perceptron (MLP) of width $2$ and depth at most $2N + 4M - 1$, with all network parameters constructed explicitly. This result is sharp with respect to width and is interpreted through the lens of simultaneous or ensemble controllability in discrete nonlinear dynamics. Second, we show that these explicit constructions yield uniform bounds on the parameter norms and, in particular, provide upper estimates for minimizers of standard regularized training loss functionals in supervised learning. As the regularization parameter vanishes, the trained networks converge to exact classifiers with bounded norm, explaining the effectiveness of overparameterized training in the small-regularization regime. We also prove a universal approximation theorem in $L^p(\Omega; \mathbb{R}_+)$ for any bounded domain $\Omega \subset \mathbb{R}^d$ and $p \in [1, \infty)$, using MLPs of fixed width $d + 1$. The proof is constructive, geometrically motivated, and provides explicit estimates on the network depth when the target function belongs to the Sobolev space $W^{1,p}$. We also extend the approximation and depth estimation results to $L^p(\Omega; \mathbb{R}^m)$ for any $m \geq 1$. Our results offer a unified and interpretable framework connecting controllability, expressivity, and training dynamics in deep neural networks.
Due to their text-to-image synthesis feature, diffusion models have recently seen a rise in visual perception tasks, such as depth estimation. The lack of good-quality datasets makes the extraction of a fine-grain semantic context challenging for the diffusion models. The semantic context with fewer details further worsens the process of creating effective text embeddings that will be used as input for diffusion models. In this paper, we propose a novel EDADepth, an enhanced data augmentation method to estimate monocular depth without using additional training data. We use Swin2SR, a super-resolution model, to enhance the quality of input images. We employ the BEiT pre-trained semantic segmentation model for better extraction of text embeddings. We use BLIP-2 tokenizer to generate tokens from these text embeddings. The novelty of our approach is the introduction of Swin2SR, the BEiT model, and the BLIP-2 tokenizer in the diffusion-based pipeline for the monocular depth estimation. Our model achieves state-of-the-art results (SOTA) on the delta3 metric on NYUv2 and KITTI datasets. It also achieves results comparable to those of the SOTA models in the RMSE and REL metrics. Finally, we also show improvements in the visualization of the estimated depth compared to the SOTA diffusion-based monocular depth estimation models. Code: https://github.com/edadepthmde/EDADepth_ICMLA.
Single-image depth estimation is essential for endoscopy tasks such as localization, reconstruction, and augmented reality. Most existing methods in surgical scenes focus on in-domain depth estimation, limiting their real-world applicability. This constraint stems from the scarcity and inferior labeling quality of medical data for training. In this work, we present EndoOmni, the first foundation model for zero-shot cross-domain depth estimation for endoscopy. To harness the potential of diverse training data, we refine the advanced self-learning paradigm that employs a teacher model to generate pseudo-labels, guiding a student model trained on large-scale labeled and unlabeled data. To address training disturbance caused by inherent noise in depth labels, we propose a robust training framework that leverages both depth labels and estimated confidence from the teacher model to jointly guide the student model training. Moreover, we propose a weighted scale-and-shift invariant loss to adaptively adjust learning weights based on label confidence, thus imposing learning bias towards cleaner label pixels while reducing the influence of highly noisy pixels. Experiments on zero-shot relative depth estimation show that our EndoOmni improves state-of-the-art methods in medical imaging for 33\% and existing foundation models for 34\% in terms of absolute relative error on specific datasets. Furthermore, our model provides strong initialization for fine-tuning metric depth estimation, maintaining superior performance in both in-domain and out-of-domain scenarios. The source code is publicly available at https://github.com/TianCuteQY/EndoOmni.
Aerial scene understanding systems face stringent payload restrictions and must often rely on monocular depth estimation for modelling scene geometry, which is an inherently ill-posed problem. Moreover, obtaining accurate ground truth data required by learning-based methods raises significant additional challenges in the aerial domain. Self-supervised approaches can bypass this problem, at the cost of providing only up-to-scale results. Similarly, recent supervised solutions which make good progress towards zero-shot generalization also provide only relative depth values. This work presents TanDepth, a practical, online scale recovery method for obtaining metric depth results from relative estimations at inference-time, irrespective of the type of model generating them. Tailored for Unmanned Aerial Vehicle (UAV) applications, our method leverages sparse measurements from Global Digital Elevation Models (GDEM) by projecting them to the camera view using extrinsic and intrinsic information. An adaptation to the Cloth Simulation Filter is presented, which allows selecting ground points from the estimated depth map to then correlate with the projected reference points. We evaluate and compare our method against alternate scaling methods adapted for UAVs, on a variety of real-world scenes. Considering the limited availability of data for this domain, we construct and release a comprehensive, depth-focused extension to the popular UAVid dataset to further research.
The increasing accuracy reports of metric monocular depth estimation models lead to a growing interest from the automotive domain. Current model evaluations do not provide deeper insights into the models' performance, also in relation to safety-critical or unseen classes. Within this paper, we present a novel approach for the evaluation of depth estimation models. Our proposed metric leverages three components, a class-wise component, an edge and corner image feature component, and a global consistency retaining component. Classes are further weighted on their distance in the scene and on criticality for automotive applications. In the evaluation, we present the benefits of our metric through comparison to classical metrics, class-wise analytics, and the retrieval of critical situations. The results show that our metric provides deeper insights into model results while fulfilling safety-critical requirements. We release the code and weights on the following repository: https://github.com/leisemann/ca_mmde