PyTorchKR Papers

최신 논문

총 2,547편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2106.12994

SGTBN: Generating Dense Depth Maps from Single-Line LiDAR

Depth completion aims to generate a dense depth map from the sparse depth map and aligned RGB image. However, current depth completion methods use extremely expensive 64-line LiDAR(about $100,000) to obtain sparse depth maps, which will limit their application scenarios. Compared with the 64-line LiDAR, the single-line LiDAR is much less expensive and much more robust. Therefore, we propose a method to tackle the problem of single-line depth completion, in which we aim to generate a dense depth map from the single-line LiDAR info and the aligned RGB image. A single-line depth completion dataset is proposed based on the existing 64-line depth completion dataset(KITTI). A network called Semantic Guided Two-Branch Network(SGTBN) which contains global and local branches to extract and fuse global and local info is proposed for this task. A Semantic guided depth upsampling module is used in our network to make full use of the semantic info in RGB images. Except for the usual MSE loss, we add the virtual normal loss to increase the constraint of high-order 3D geometry in our network. Our network outperforms the state-of-the-art in the single-line depth completion task. Besides, compared with the monocular depth estimation, our method also has significant advantages in precision and model size.

arXiv:2106.12958공식 코드 · Papers with Code

Self-Supervised Monocular Depth Estimation of Untextured Indoor Rotated Scenes

Self-supervised deep learning methods have leveraged stereo images for training monocular depth estimation. Although these methods show strong results on outdoor datasets such as KITTI, they do not match performance of supervised methods on indoor environments with camera rotation. Indoor, rotated scenes are common for less constrained applications and pose problems for two reasons: abundance of low texture regions and increased complexity of depth cues for images under rotation. In an effort to extend self-supervised learning to more generalised environments we propose two additions. First, we propose a novel Filled Disparity Loss term that corrects for ambiguity of image reconstruction error loss in textureless regions. Specifically, we interpolate disparity in untextured regions, using the estimated disparity from surrounding textured areas, and use L1 loss to correct the original estimation. Our experiments show that depth estimation is substantially improved on low-texture scenes, without any loss on textured scenes, when compared to Monodepth by Godard et al. Secondly, we show that training with an application's representative rotations, in both pitch and roll, is sufficient to significantly improve performance over the entire range of expected rotation. We demonstrate that depth estimation is successfully generalised as performance is not lost when evaluated on test sets with no camera rotation. Together these developments enable a broader use of self-supervised learning of monocular depth estimation for complex environments.

arXiv:2106.08615

EdgeConv with Attention Module for Monocular Depth Estimation

Monocular depth estimation is an especially important task in robotics and autonomous driving, where 3D structural information is essential. However, extreme lighting conditions and complex surface objects make it difficult to predict depth in a single image. Therefore, to generate accurate depth maps, it is important for the model to learn structural information about the scene. We propose a novel Patch-Wise EdgeConv Module (PEM) and EdgeConv Attention Module (EAM) to solve the difficulty of monocular depth estimation. The proposed modules extract structural information by learning the relationship between image patches close to each other in space using edge convolution. Our method is evaluated on two popular datasets, the NYU Depth V2 and the KITTI Eigen split, achieving state-of-the-art performance. We prove that the proposed model predicts depth robustly in challenging scenes through various comparative experiments.

arXiv:2106.08486

Achieving Domain Robustness in Stereo Matching Networks by Removing Shortcut Learning

Learning-based stereo matching and depth estimation networks currently excel on public benchmarks with impressive results. However, state-of-the-art networks often fail to generalize from synthetic imagery to more challenging real data domains. This paper is an attempt to uncover hidden secrets of achieving domain robustness and in particular, discovering the important ingredients of generalization success of stereo matching networks by analyzing the effect of synthetic image learning on real data performance. We provide evidence that demonstrates that learning of features in the synthetic domain by a stereo matching network is heavily influenced by two "shortcuts" presented in the synthetic data: (1) identical local statistics (RGB colour features) between matching pixels in the synthetic stereo images and (2) lack of realism in synthetic textures on 3D objects simulated in game engines. We will show that by removing such shortcuts, we can achieve domain robustness in the state-of-the-art stereo matching frameworks and produce a remarkable performance on multiple realistic datasets, despite the fact that the networks were trained on synthetic data, only. Our experimental results point to the fact that eliminating shortcuts from the synthetic data is key to achieve domain-invariant generalization between synthetic and real data domains.

arXiv:2106.08917공식 코드 · Papers with Code

Differentiable Diffusion for Dense Depth Estimation from Multi-view Images

We present a method to estimate dense depth by optimizing a sparse set of points such that their diffusion into a depth map minimizes a multi-view reprojection error from RGB supervision. We optimize point positions, depths, and weights with respect to the loss by differential splatting that models points as Gaussians with analytic transmittance. Further, we develop an efficient optimization routine that can simultaneously optimize the 50k+ points required for complex scene reconstruction. We validate our routine using ground truth data and show high reconstruction quality. Then, we apply this to light field and wider baseline images via self supervision, and show improvements in both average and outlier error for depth maps diffused from inaccurate sparse points. Finally, we compare qualitative and quantitative results to image processing and deep learning methods. http://visual.cs.brown.edu/diffdiffdepth

arXiv:2106.08320공식 코드 · Papers with Code

Self-Supervised Learning with Kernel Dependence Maximization

We approach self-supervised learning of image representations from a statistical dependence perspective, proposing Self-Supervised Learning with the Hilbert-Schmidt Independence Criterion (SSL-HSIC). SSL-HSIC maximizes dependence between representations of transformations of an image and the image identity, while minimizing the kernelized variance of those representations. This framework yields a new understanding of InfoNCE, a variational lower bound on the mutual information (MI) between different transformations. While the MI itself is known to have pathologies which can result in learning meaningless representations, its bound is much better behaved: we show that it implicitly approximates SSL-HSIC (with a slightly different regularizer). Our approach also gives us insight into BYOL, a negative-free SSL method, since SSL-HSIC similarly learns local neighborhoods of samples. SSL-HSIC allows us to directly optimize statistical dependence in time linear in the batch size, without restrictive data assumptions or indirect mutual information estimators. Trained with or without a target network, SSL-HSIC matches the current state-of-the-art for standard linear evaluation on ImageNet, semi-supervised learning and transfer to other classification and vision tasks such as semantic segmentation, depth estimation and object recognition. Code is available at https://github.com/deepmind/ssl_hsic .

arXiv:2106.07856

A Hybrid mmWave and Camera System for Long-Range Depth Imaging

mmWave radars offer excellent depth resolution even at very long ranges owing to their high bandwidth. But their angular resolution is at least an order-of-magnitude worse than camera and lidar systems. Hence, mmWave radar is not a capable 3-D imaging solution in isolation. We propose Metamoran, a system that combines the complimentary strengths of radar and camera to obtain accurate, high resolution depth images over long ranges even in high clutter environments, all from a single fixed vantage point. Metamoran enables rich long-range depth imaging with applications in security and surveillance, roadside safety infrastructure and wide-area mapping. Our approach leverages the high angular resolution from cameras using computer vision techniques, including image segmentation and monocular depth estimation, to obtain object shape. Our core contribution is a method to convert this object shape into an RF I/Q equivalent, which we use in a novel radar processing pipeline to help declutter the scene and capture extremely weak reflections from objects at long distances. We perform a detailed evaluation of Metamoran's depth imaging capabilities in 400 diverse scenes. Our evaluation shows that Metamoran estimates the depth of static objects up to 90 m and moving objects up to 305 m and with a median error of 28 cm, an improvement of 13$\times$ compared to a naive radar+camera baseline and 23$\times$ compared to monocular depth estimation.

arXiv:2106.06629공식 코드 · Papers with Code

Mirror3D: Depth Refinement for Mirror Surfaces

Despite recent progress in depth sensing and 3D reconstruction, mirror surfaces are a significant source of errors. To address this problem, we create the Mirror3D dataset: a 3D mirror plane dataset based on three RGBD datasets (Matterport3D, NYUv2 and ScanNet) containing 7,011 mirror instance masks and 3D planes. We then develop Mirror3DNet: a module that refines raw sensor depth or estimated depth to correct errors on mirror surfaces. Our key idea is to estimate the 3D mirror plane based on RGB input and surrounding depth context, and use this estimate to directly regress mirror surface depth. Our experiments show that Mirror3DNet significantly mitigates errors from a variety of input depth data, including raw sensor depth and depth estimation or completion methods.

arXiv:2106.07256

Deterministic Guided LiDAR Depth Map Completion

Accurate dense depth estimation is crucial for autonomous vehicles to analyze their environment. This paper presents a non-deep learning-based approach to densify a sparse LiDAR-based depth map using a guidance RGB image. To achieve this goal the RGB image is at first cleared from most of the camera-LiDAR misalignment artifacts. Afterward, it is over segmented and a plane for each superpixel is approximated. In the case a superpixel is not well represented by a plane, a plane is approximated for a convex hull of the most inlier. Finally, the pinhole camera model is used for the interpolation process and the remaining areas are interpolated. The evaluation of this work is executed using the KITTI depth completion benchmark, which validates the proposed work and shows that it outperforms the state-of-the-art non-deep learning-based methods, in addition to several deep learning-based methods.

arXiv:2106.06129

Instance-Level Task Parameters: A Robust Multi-task Weighting Framework

Recent works have shown that deep neural networks benefit from multi-task learning by learning a shared representation across several related tasks. However, performance of such systems depend on relative weighting between various losses involved during training. Prior works on loss weighting schemes assume that instances are equally easy or hard for all tasks. In order to break this assumption, we let the training process dictate the optimal weighting of tasks for every instance in the dataset. More specifically, we equip every instance in the dataset with a set of learnable parameters (instance-level task parameters) where the cardinality is equal to the number of tasks learned by the model. These parameters model the weighting of each task for an instance. They are updated by gradient descent and do not require hand-crafted rules. We conduct extensive experiments on SURREAL and CityScapes datasets, for human shape and pose estimation, depth estimation and semantic segmentation tasks. In these tasks, our approach outperforms recent dynamic loss weighting approaches, e.g. reducing surface estimation errors by 8.97% on SURREAL. When applied to datasets where one or more tasks can have noisy annotations, the proposed method learns to prioritize learning from clean labels for a given task, e.g. reducing surface estimation errors by up to 60%. We also show that we can reliably detect corrupt labels for a given task as a by-product from learned instance-level task parameters.

arXiv:2106.05525

3D Semantic Mapping from Arthroscopy using Out-of-distribution Pose and Depth and In-distribution Segmentation Training

Minimally invasive surgery (MIS) has many documented advantages, but the surgeon's limited visual contact with the scene can be problematic. Hence, systems that can help surgeons navigate, such as a method that can produce a 3D semantic map, can compensate for the limitation above. In theory, we can borrow 3D semantic mapping techniques developed for robotics, but this requires finding solutions to the following challenges in MIS: 1) semantic segmentation, 2) depth estimation, and 3) pose estimation. In this paper, we propose the first 3D semantic mapping system from knee arthroscopy that solves the three challenges above. Using out-of-distribution non-human datasets, where pose could be labeled, we jointly train depth+pose estimators using selfsupervised and supervised losses. Using an in-distribution human knee dataset, we train a fully-supervised semantic segmentation system to label arthroscopic image pixels into femur, ACL, and meniscus. Taking testing images from human knees, we combine the results from these two systems to automatically create 3D semantic maps of the human knee. The result of this work opens the pathway to the generation of intraoperative 3D semantic mapping, registration with pre-operative data, and robotic-assisted arthroscopy

arXiv:2106.03043공식 코드 · Papers with Code

Occlusion-aware Unsupervised Learning of Depth from 4-D Light Fields

Depth estimation is a fundamental issue in 4-D light field processing and analysis. Although recent supervised learning-based light field depth estimation methods have significantly improved the accuracy and efficiency of traditional optimization-based ones, these methods rely on the training over light field data with ground-truth depth maps which are challenging to obtain or even unavailable for real-world light field data. Besides, due to the inevitable gap (or domain difference) between real-world and synthetic data, they may suffer from serious performance degradation when generalizing the models trained with synthetic data to real-world data. By contrast, we propose an unsupervised learning-based method, which does not require ground-truth depth as supervision during training. Specifically, based on the basic knowledge of the unique geometry structure of light field data, we present an occlusion-aware strategy to improve the accuracy on occlusion areas, in which we explore the angular coherence among subsets of the light field views to estimate initial depth maps, and utilize a constrained unsupervised loss to learn their corresponding reliability for final depth prediction. Additionally, we adopt a multi-scale network with a weighted smoothness loss to handle the textureless areas. Experimental results on synthetic data show that our method can significantly shrink the performance gap between the previous unsupervised method and supervised ones, and produce depth maps with comparable accuracy to traditional methods with obviously reduced computational cost. Moreover, experiments on real-world datasets show that our method can avoid the domain shift problem presented in supervised methods, demonstrating the great potential of our method.

arXiv:2106.03505

Self-supervised Depth Estimation Leveraging Global Perception and Geometric Smoothness Using On-board Videos

Self-supervised depth estimation has drawn much attention in recent years as it does not require labeled data but image sequences. Moreover, it can be conveniently used in various applications, such as autonomous driving, robotics, realistic navigation, and smart cities. However, extracting global contextual information from images and predicting a geometrically natural depth map remain challenging. In this paper, we present DLNet for pixel-wise depth estimation, which simultaneously extracts global and local features with the aid of our depth Linformer block. This block consists of the Linformer and innovative soft split multi-layer perceptron blocks. Moreover, a three-dimensional geometry smoothness loss is proposed to predict a geometrically natural depth map by imposing the second-order smoothness constraint on the predicted three-dimensional point clouds, thereby realizing improved performance as a byproduct. Finally, we explore the multi-scale prediction strategy and propose the maximum margin dual-scale prediction strategy for further performance improvement. In experiments on the KITTI and Make3D benchmarks, the proposed DLNet achieves performance competitive to those of the state-of-the-art methods, reducing time and space complexities by more than $62\%$ and $56\%$, respectively. Extensive testing on various real-world situations further demonstrates the strong practicality and generalization capability of the proposed model.

arXiv:2106.02594

Self-Supervised Learning of Domain Invariant Features for Depth Estimation

We tackle the problem of unsupervised synthetic-to-real domain adaptation for single image depth estimation. An essential building block of single image depth estimation is an encoder-decoder task network that takes RGB images as input and produces depth maps as output. In this paper, we propose a novel training strategy to force the task network to learn domain invariant representations in a selfsupervised manner. Specifically, we extend self-supervised learning from traditional representation learning, which works on images from a single domain, to domain invariant representation learning, which works on images from two different domains by utilizing an image-to-image translation network. Firstly, we use an image-to-image translation network to transfer domain-specific styles between synthetic and real domains. This style transfer operation allows us to obtain similar images from the different domains. Secondly, we jointly train our task network and Siamese network with the same images from the different domains to obtain domain invariance for the task network. Finally, we fine-tune the task network using labeled synthetic and unlabeled realworld data. Our training strategy yields improved generalization capability in the real-world domain. We carry out an extensive evaluation on two popular datasets for depth estimation, KITTI and Make3D. The results demonstrate that our proposed method outperforms the state-of-the-art on all metrics, e.g. by 14.7% on Sq Rel on KITTI. The source code and model weights will be made available.

arXiv:2106.01717

Towards urban scenes understanding through polarization cues

Autonomous robotics is critically affected by the robustness of its scene understanding algorithms. We propose a two-axis pipeline based on polarization indices to analyze dynamic urban scenes. As robots evolve in unknown environments, they are prone to encountering specular obstacles. Usually, specular phenomena are rarely taken into account by algorithms which causes misinterpretations and erroneous estimates. By exploiting all the light properties, systems can greatly increase their robustness to events. In addition to the conventional photometric characteristics, we propose to include polarization sensing. We demonstrate in this paper that the contribution of polarization measurement increases both the performances of segmentation and the quality of depth estimation. Our polarimetry-based approaches are compared here with other state-of-the-art RGB-centric methods showing interest of using polarization imaging.

arXiv:2106.02022공식 코드 · Papers with Code

Single Image Depth Prediction with Wavelet Decomposition

We present a novel method for predicting accurate depths from monocular images with high efficiency. This optimal efficiency is achieved by exploiting wavelet decomposition, which is integrated in a fully differentiable encoder-decoder architecture. We demonstrate that we can reconstruct high-fidelity depth maps by predicting sparse wavelet coefficients. In contrast with previous works, we show that wavelet coefficients can be learned without direct supervision on coefficients. Instead we supervise only the final depth image that is reconstructed through the inverse wavelet transform. We additionally show that wavelet coefficients can be learned in fully self-supervised scenarios, without access to ground-truth depth. Finally, we apply our method to different state-of-the-art monocular depth estimation models, in each case giving similar or better results compared to the original model, while requiring less than half the multiply-adds in the decoder network. Code at https://github.com/nianticlabs/wavelet-monodepth

arXiv:2105.14520공식 코드 · Papers with Code

Unsupervised Joint Learning of Depth, Optical Flow, Ego-motion from Video

Estimating geometric elements such as depth, camera motion, and optical flow from images is an important part of the robot's visual perception. We use a joint self-supervised method to estimate the three geometric elements. Depth network, optical flow network and camera motion network are independent of each other but are jointly optimized during training phase. Compared with independent training, joint training can make full use of the geometric relationship between geometric elements and provide dynamic and static information of the scene. In this paper, we improve the joint self-supervision method from three aspects: network structure, dynamic object segmentation, and geometric constraints. In terms of network structure, we apply the attention mechanism to the camera motion network, which helps to take advantage of the similarity of camera movement between frames. And according to attention mechanism in Transformer, we propose a plug-and-play convolutional attention module. In terms of dynamic object, according to the different influences of dynamic objects in the optical flow self-supervised framework and the depth-pose self-supervised framework, we propose a threshold algorithm to detect dynamic regions, and mask that in the loss function respectively. In terms of geometric constraints, we use traditional methods to estimate the fundamental matrix from the corresponding points to constrain the camera motion network. We demonstrate the effectiveness of our method on the KITTI dataset. Compared with other joint self-supervised methods, our method achieves state-of-the-art performance in the estimation of pose and optical flow, and the depth estimation has also achieved competitive results. Code will be available https://github.com/jianfenglihg/Unsupervised_geometry.

arXiv:2105.14021공식 코드 · Papers with Code

Boosting Monocular Depth Estimation Models to High-Resolution via Content-Adaptive Multi-Resolution Merging

Neural networks have shown great abilities in estimating depth from a single image. However, the inferred depth maps are well below one-megapixel resolution and often lack fine-grained details, which limits their practicality. Our method builds on our analysis on how the input resolution and the scene structure affects depth estimation performance. We demonstrate that there is a trade-off between a consistent scene structure and the high-frequency details, and merge low- and high-resolution estimations to take advantage of this duality using a simple depth merging network. We present a double estimation method that improves the whole-image depth estimation and a patch selection method that adds local details to the final result. We demonstrate that by merging estimations at different resolutions with changing context, we can generate multi-megapixel depth maps with a high level of detail using a pre-trained model.

arXiv:2105.12186

Self-Guided Instance-Aware Network for Depth Completion and Enhancement

Depth completion aims at inferring a dense depth image from sparse depth measurement since glossy, transparent or distant surface cannot be scanned properly by the sensor. Most of existing methods directly interpolate the missing depth measurements based on pixel-wise image content and the corresponding neighboring depth values. Consequently, this leads to blurred boundaries or inaccurate structure of object. To address these problems, we propose a novel self-guided instance-aware network (SG-IANet) that: (1) utilize self-guided mechanism to extract instance-level features that is needed for depth restoration, (2) exploit the geometric and context information into network learning to conform to the underlying constraints for edge clarity and structure consistency, (3) regularize the depth estimation and mitigate the impact of noise by instance-aware learning, and (4) train with synthetic data only by domain randomization to bridge the reality gap. Extensive experiments on synthetic and real world dataset demonstrate that our proposed method outperforms previous works. Further ablation studies give more insights into the proposed method and demonstrate the generalization capability of our model.

arXiv:2105.12053

Real-time Monocular Depth Estimation with Sparse Supervision on Mobile

Monocular (relative or metric) depth estimation is a critical task for various applications, such as autonomous vehicles, augmented reality and image editing. In recent years, with the increasing availability of mobile devices, accurate and mobile-friendly depth models have gained importance. Increasingly accurate models typically require more computational resources, which inhibits the use of such models on mobile devices. The mobile use case is arguably the most unrestricted one, which requires highly accurate yet mobile-friendly architectures. Therefore, we try to answer the following question: How can we improve a model without adding further complexity (i.e. parameters)? Towards this end, we systematically explore the design space of a relative depth estimation model from various dimensions and we show, with key design choices and ablation studies, even an existing architecture can reach highly competitive performance to the state of the art, with a fraction of the complexity. Our study spans an in-depth backbone model selection process, knowledge distillation, intermediate predictions, model pruning and loss rebalancing. We show that our model, using only DIW as the supervisory dataset, achieves 0.1156 WHDR on DIW with 2.6M parameters and reaches 37 FPS on a mobile GPU, without pruning or hardware-specific optimization. A pruned version of our model achieves 0.1208 WHDR on DIW with 1M parameters and reaches 44 FPS on a mobile GPU.