Self-supervised depth estimation draws a lot of attention recently as it can promote the 3D sensing capabilities of self-driving vehicles. However, it intrinsically relies upon the photometric consistency assumption, which hardly holds during nighttime. Although various supervised nighttime image enhancement methods have been proposed, their generalization performance in challenging driving scenarios is not satisfactory. To this end, we propose the first method that jointly learns a nighttime image enhancer and a depth estimator, without using ground truth for either task. Our method tightly entangles two self-supervised tasks using a newly proposed uncertain pixel masking strategy. This strategy originates from the observation that nighttime images not only suffer from underexposed regions but also from overexposed regions. By fitting a bridge-shaped curve to the illumination map distribution, both regions are suppressed and two tasks are bridged naturally. We benchmark the method on two established datasets: nuScenes and RobotCar and demonstrate state-of-the-art performance on both of them. Detailed ablations also reveal the mechanism of our proposal. Last but not least, to mitigate the problem of sparse ground truth of existing datasets, we provide a new photo-realistically enhanced nighttime dataset based upon CARLA. It brings meaningful new challenges to the community. Codes, data, and models are available at //github.com/ucaszyp/STEPS.
相關內容
Multi-contrast magnetic resonance imaging (MRI) is the most common management tool used to characterize neurological disorders based on brain tissue contrasts. However, acquiring high-resolution MRI scans is time-consuming and infeasible under specific conditions. Hence, multi-contrast super-resolution methods have been developed to improve the quality of low-resolution contrasts by leveraging complementary information from multi-contrast MRI. Current deep learning-based super-resolution methods have limitations in estimating restoration uncertainty and avoiding mode collapse. Although the diffusion model has emerged as a promising approach for image enhancement, capturing complex interactions between multiple conditions introduced by multi-contrast MRI super-resolution remains a challenge for clinical applications. In this paper, we propose a disentangled conditional diffusion model, DisC-Diff, for multi-contrast brain MRI super-resolution. It utilizes the sampling-based generation and simple objective function of diffusion models to estimate uncertainty in restorations effectively and ensure a stable optimization process. Moreover, DisC-Diff leverages a disentangled multi-stream network to fully exploit complementary information from multi-contrast MRI, improving model interpretation under multiple conditions of multi-contrast inputs. We validated the effectiveness of DisC-Diff on two datasets: the IXI dataset, which contains 578 normal brains, and a clinical dataset with 316 pathological brains. Our experimental results demonstrate that DisC-Diff outperforms other state-of-the-art methods both quantitatively and visually.
Joint radar-communications (JRC) has emerged as a promising technology for efficiently using the limited electromagnetic spectrum. In JRC applications such as secure military receivers, often the radar and communications signals are overlaid in the received signal. In these passive listening outposts, the signals and channels of both radar and communications are unknown to the receiver. The ill-posed problem of recovering all signal and channel parameters from the overlaid signal is terms as dual-blind deconvolution (DBD). In this work, we investigate a more challenging version of DBD with a multi-antenna receiver. We model the radar and communications channels with a few (sparse) continuous-valued parameters such as time delays, Doppler velocities, and directions-of-arrival (DoAs). To solve this highly ill-posed DBD, we propose to minimize the sum of multivariate atomic norms (SoMAN) that depends on the unknown parameters. To this end, we devise an exact semidefinite program using theories of positive hyperoctant trigonometric polynomials (PhTP). Our theoretical analyses show that the minimum number of samples and antennas required for perfect recovery is logarithmically dependent on the maximum of the number of radar targets and communications paths rather than their sum. We show that our approach is easily generalized to include several practical issues such as gain/phase errors and additive noise. Numerical experiments show the exact parameter recovery for different JRC
In many automation tasks involving manipulation of rigid objects, the poses of the objects must be acquired. Vision-based pose estimation using a single RGB or RGB-D sensor is especially popular due to its broad applicability. However, single-view pose estimation is inherently limited by depth ambiguity and ambiguities imposed by various phenomena like occlusion, self-occlusion, reflections, etc. Aggregation of information from multiple views can potentially resolve these ambiguities, but the current state-of-the-art multi-view pose estimation method only uses multiple views to aggregate single-view pose estimates, and thus rely on obtaining good single-view estimates. We present a multi-view pose estimation method which aggregates learned 2D-3D distributions from multiple views for both the initial estimate and optional refinement. Our method performs probabilistic sampling of 3D-3D correspondences under epipolar constraints using learned 2D-3D correspondence distributions which are implicitly trained to respect visual ambiguities such as symmetry. Evaluation on the T-LESS dataset shows that our method reduces pose estimation errors by 80-91% compared to the best single-view method, and we present state-of-the-art results on T-LESS with four views, even compared with methods using five and eight views.
Existing methods for 3D-aware image synthesis largely depend on the 3D pose distribution pre-estimated on the training set. An inaccurate estimation may mislead the model into learning faulty geometry. This work proposes PoF3D that frees generative radiance fields from the requirements of 3D pose priors. We first equip the generator with an efficient pose learner, which is able to infer a pose from a latent code, to approximate the underlying true pose distribution automatically. We then assign the discriminator a task to learn pose distribution under the supervision of the generator and to differentiate real and synthesized images with the predicted pose as the condition. The pose-free generator and the pose-aware discriminator are jointly trained in an adversarial manner. Extensive results on a couple of datasets confirm that the performance of our approach, regarding both image quality and geometry quality, is on par with state of the art. To our best knowledge, PoF3D demonstrates the feasibility of learning high-quality 3D-aware image synthesis without using 3D pose priors for the first time.
Multiple near frontal-parallel planes based depth representation demonstrated impressive results in self-supervised monocular depth estimation (MDE). Whereas, such a representation would cause the discontinuity of the ground as it is perpendicular to the frontal-parallel planes, which is detrimental to the identification of drivable space in autonomous driving. In this paper, we propose the PlaneDepth, a novel orthogonal planes based presentation, including vertical planes and ground planes. PlaneDepth estimates the depth distribution using a Laplacian Mixture Model based on orthogonal planes for an input image. These planes are used to synthesize a reference view to provide the self-supervision signal. Further, we find that the widely used resizing and cropping data augmentation breaks the orthogonality assumptions, leading to inferior plane predictions. We address this problem by explicitly constructing the resizing cropping transformation to rectify the predefined planes and predicted camera pose. Moreover, we propose an augmented self-distillation loss supervised with a bilateral occlusion mask to boost the robustness of orthogonal planes representation for occlusions. Thanks to our orthogonal planes representation, we can extract the ground plane in an unsupervised manner, which is important for autonomous driving. Extensive experiments on the KITTI dataset demonstrate the effectiveness and efficiency of our method. The code is available at //github.com/svip-lab/PlaneDepth.
In this study, we present a method for synthesizing novel views from a single 360-degree RGB-D image based on the neural radiance field (NeRF) . Prior studies relied on the neighborhood interpolation capability of multi-layer perceptrons to complete missing regions caused by occlusion and zooming, which leads to artifacts. In the method proposed in this study, the input image is reprojected to 360-degree RGB images at other camera positions, the missing regions of the reprojected images are completed by a 2D image generative model, and the completed images are utilized to train the NeRF. Because multiple completed images contain inconsistencies in 3D, we introduce a method to learn the NeRF model using a subset of completed images that cover the target scene with less overlap of completed regions. The selection of such a subset of images can be attributed to the maximum weight independent set problem, which is solved through simulated annealing. Experiments demonstrated that the proposed method can synthesize plausible novel views while preserving the features of the scene for both artificial and real-world data.
When handling complicated text images (e.g., irregular structures, low resolution, heavy occlusion, and uneven illumination), existing supervised text recognition methods are data-hungry. Although these methods employ large-scale synthetic text images to reduce the dependence on annotated real images, the domain gap still limits the recognition performance. Therefore, exploring the robust text feature representations on unlabeled real images by self-supervised learning is a good solution. However, existing self-supervised text recognition methods conduct sequence-to-sequence representation learning by roughly splitting the visual features along the horizontal axis, which limits the flexibility of the augmentations, as large geometric-based augmentations may lead to sequence-to-sequence feature inconsistency. Motivated by this, we propose a novel self-supervised Character-to-Character Distillation method, CCD, which enables versatile augmentations to facilitate general text representation learning. Specifically, we delineate the character structures of unlabeled real images by designing a self-supervised character segmentation module. Following this, CCD easily enriches the diversity of local characters while keeping their pairwise alignment under flexible augmentations, using the transformation matrix between two augmented views from images. Experiments demonstrate that CCD achieves state-of-the-art results, with average performance gains of 1.38% in text recognition, 1.7% in text segmentation, 0.24 dB (PSNR) and 0.0321 (SSIM) in text super-resolution. Code will be released soon.
Depth-from-defocus (DFD), modeling the relationship between depth and defocus pattern in images, has demonstrated promising performance in depth estimation. Recently, several self-supervised works try to overcome the difficulties in acquiring accurate depth ground-truth. However, they depend on the all-in-focus (AIF) images, which cannot be captured in real-world scenarios. Such limitation discourages the applications of DFD methods. To tackle this issue, we propose a completely self-supervised framework that estimates depth purely from a sparse focal stack. We show that our framework circumvents the needs for the depth and AIF image ground-truth, and receives superior predictions, thus closing the gap between the theoretical success of DFD works and their applications in the real world. In particular, we propose (i) a more realistic setting for DFD tasks, where no depth or AIF image ground-truth is available; (ii) a novel self-supervision framework that provides reliable predictions of depth and AIF image under the challenging setting. The proposed framework uses a neural model to predict the depth and AIF image, and utilizes an optical model to validate and refine the prediction. We verify our framework on three benchmark datasets with rendered focal stacks and real focal stacks. Qualitative and quantitative evaluations show that our method provides a strong baseline for self-supervised DFD tasks.
Few-shot image classification aims to classify unseen classes with limited labeled samples. Recent works benefit from the meta-learning process with episodic tasks and can fast adapt to class from training to testing. Due to the limited number of samples for each task, the initial embedding network for meta learning becomes an essential component and can largely affects the performance in practice. To this end, many pre-trained methods have been proposed, and most of them are trained in supervised way with limited transfer ability for unseen classes. In this paper, we proposed to train a more generalized embedding network with self-supervised learning (SSL) which can provide slow and robust representation for downstream tasks by learning from the data itself. We evaluate our work by extensive comparisons with previous baseline methods on two few-shot classification datasets ({\em i.e.,} MiniImageNet and CUB). Based on the evaluation results, the proposed method achieves significantly better performance, i.e., improve 1-shot and 5-shot tasks by nearly \textbf{3\%} and \textbf{4\%} on MiniImageNet, by nearly \textbf{9\%} and \textbf{3\%} on CUB. Moreover, the proposed method can gain the improvement of (\textbf{15\%}, \textbf{13\%}) on MiniImageNet and (\textbf{15\%}, \textbf{8\%}) on CUB by pretraining using more unlabeled data. Our code will be available at \hyperref[//github.com/phecy/SSL-FEW-SHOT.]{//github.com/phecy/ssl-few-shot.}
This work addresses a novel and challenging problem of estimating the full 3D hand shape and pose from a single RGB image. Most current methods in 3D hand analysis from monocular RGB images only focus on estimating the 3D locations of hand keypoints, which cannot fully express the 3D shape of hand. In contrast, we propose a Graph Convolutional Neural Network (Graph CNN) based method to reconstruct a full 3D mesh of hand surface that contains richer information of both 3D hand shape and pose. To train networks with full supervision, we create a large-scale synthetic dataset containing both ground truth 3D meshes and 3D poses. When fine-tuning the networks on real-world datasets without 3D ground truth, we propose a weakly-supervised approach by leveraging the depth map as a weak supervision in training. Through extensive evaluations on our proposed new datasets and two public datasets, we show that our proposed method can produce accurate and reasonable 3D hand mesh, and can achieve superior 3D hand pose estimation accuracy when compared with state-of-the-art methods.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191