We propose SUB-Depth, a universal multi-task training framework for self-supervised monocular depth estimation (SDE). Depth models trained with SUB-Depth outperform the same models trained in a standard single-task SDE framework. By introducing an additional self-distillation task into a standard SDE training framework, SUB-Depth trains a depth network, not only to predict the depth map for an image reconstruction task, but also to distill knowledge from a trained teacher network with unlabelled data. To take advantage of this multi-task setting, we propose homoscedastic uncertainty formulations for each task to penalize areas likely to be affected by teacher network noise, or violate SDE assumptions. We present extensive evaluations on KITTI to demonstrate the improvements achieved by training a range of existing networks using the proposed framework, and we achieve state-of-the-art performance on this task. Additionally, SUB-Depth enables models to estimate uncertainty on depth output.
相關內容
Most change detection methods assume that pre-change and post-change images are acquired by the same sensor. However, in many real-life scenarios, e.g., natural disaster, it is more practical to use the latest available images before and after the occurrence of incidence, which may be acquired using different sensors. In particular, we are interested in the combination of the images acquired by optical and Synthetic Aperture Radar (SAR) sensors. SAR images appear vastly different from the optical images even when capturing the same scene. Adding to this, change detection methods are often constrained to use only target image-pair, no labeled data, and no additional unlabeled data. Such constraints limit the scope of traditional supervised machine learning and unsupervised generative approaches for multi-sensor change detection. Recent rapid development of self-supervised learning methods has shown that some of them can even work with only few images. Motivated by this, in this work we propose a method for multi-sensor change detection using only the unlabeled target bi-temporal images that are used for training a network in self-supervised fashion by using deep clustering and contrastive learning. The proposed method is evaluated on four multi-modal bi-temporal scenes showing change and the benefits of our self-supervised approach are demonstrated.
3D reconstruction of depth and motion from monocular video in dynamic environments is a highly ill-posed problem due to scale ambiguities when projecting to the 2D image domain. In this work, we investigate the performance of the current State-of-the-Art (SotA) deep multi-view systems in such environments. We find that current supervised methods work surprisingly well despite not modelling individual object motions, but make systematic errors due to a lack of dense ground truth data. To detect such errors during usage, we extend the cost volume based Deep Video to Depth (DeepV2D) framework \cite{teed2018deepv2d} with a learned uncertainty. Our Deep Video to certain Depth (DeepV2cD) model allows i) to perform en par or better with current SotA and ii) achieve a better uncertainty measure than the naive Shannon entropy. Our experiments show that a simple filter strategy based on the uncertainty can significantly reduce systematic errors. This results in cleaner reconstructions both on static and dynamic parts of the scene.
Transformers are state-of-the-art in a wide range of NLP tasks and have also been applied to many real-world products. Understanding the reliability and certainty of transformer model predictions is crucial for building trustable machine learning applications, e.g., medical diagnosis. Although many recent transformer extensions have been proposed, the study of the uncertainty estimation of transformer models is under-explored. In this work, we propose a novel way to enable transformers to have the capability of uncertainty estimation and, meanwhile, retain the original predictive performance. This is achieved by learning a hierarchical stochastic self-attention that attends to values and a set of learnable centroids, respectively. Then new attention heads are formed with a mixture of sampled centroids using the Gumbel-Softmax trick. We theoretically show that the self-attention approximation by sampling from a Gumbel distribution is upper bounded. We empirically evaluate our model on two text classification tasks with both in-domain (ID) and out-of-domain (OOD) datasets. The experimental results demonstrate that our approach: (1) achieves the best predictive performance and uncertainty trade-off among compared methods; (2) exhibits very competitive (in most cases, improved) predictive performance on ID datasets; (3) is on par with Monte Carlo dropout and ensemble methods in uncertainty estimation on OOD datasets.
A video autoencoder is proposed for learning disentan- gled representations of 3D structure and camera pose from videos in a self-supervised manner. Relying on temporal continuity in videos, our work assumes that the 3D scene structure in nearby video frames remains static. Given a sequence of video frames as input, the video autoencoder extracts a disentangled representation of the scene includ- ing: (i) a temporally-consistent deep voxel feature to represent the 3D structure and (ii) a 3D trajectory of camera pose for each frame. These two representations will then be re-entangled for rendering the input video frames. This video autoencoder can be trained directly using a pixel reconstruction loss, without any ground truth 3D or camera pose annotations. The disentangled representation can be applied to a range of tasks, including novel view synthesis, camera pose estimation, and video generation by motion following. We evaluate our method on several large- scale natural video datasets, and show generalization results on out-of-domain images.
Remarkable results have been achieved by DCNN based self-supervised depth estimation approaches. However, most of these approaches can only handle either day-time or night-time images, while their performance degrades for all-day images due to large domain shift and the variation of illumination between day and night images. To relieve these limitations, we propose a domain-separated network for self-supervised depth estimation of all-day images. Specifically, to relieve the negative influence of disturbing terms (illumination, etc.), we partition the information of day and night image pairs into two complementary sub-spaces: private and invariant domains, where the former contains the unique information (illumination, etc.) of day and night images and the latter contains essential shared information (texture, etc.). Meanwhile, to guarantee that the day and night images contain the same information, the domain-separated network takes the day-time images and corresponding night-time images (generated by GAN) as input, and the private and invariant feature extractors are learned by orthogonality and similarity loss, where the domain gap can be alleviated, thus better depth maps can be expected. Meanwhile, the reconstruction and photometric losses are utilized to estimate complementary information and depth maps effectively. Experimental results demonstrate that our approach achieves state-of-the-art depth estimation results for all-day images on the challenging Oxford RobotCar dataset, proving the superiority of our proposed approach.
Heatmap-based methods dominate in the field of human pose estimation by modelling the output distribution through likelihood heatmaps. In contrast, regression-based methods are more efficient but suffer from inferior performance. In this work, we explore maximum likelihood estimation (MLE) to develop an efficient and effective regression-based methods. From the perspective of MLE, adopting different regression losses is making different assumptions about the output density function. A density function closer to the true distribution leads to a better regression performance. In light of this, we propose a novel regression paradigm with Residual Log-likelihood Estimation (RLE) to capture the underlying output distribution. Concretely, RLE learns the change of the distribution instead of the unreferenced underlying distribution to facilitate the training process. With the proposed reparameterization design, our method is compatible with off-the-shelf flow models. The proposed method is effective, efficient and flexible. We show its potential in various human pose estimation tasks with comprehensive experiments. Compared to the conventional regression paradigm, regression with RLE bring 12.4 mAP improvement on MSCOCO without any test-time overhead. Moreover, for the first time, especially on multi-person pose estimation, our regression method is superior to the heatmap-based methods. Our code is available at //github.com/Jeff-sjtu/res-loglikelihood-regression
To date, most existing self-supervised learning methods are designed and optimized for image classification. These pre-trained models can be sub-optimal for dense prediction tasks due to the discrepancy between image-level prediction and pixel-level prediction. To fill this gap, we aim to design an effective, dense self-supervised learning method that directly works at the level of pixels (or local features) by taking into account the correspondence between local features. We present dense contrastive learning, which implements self-supervised learning by optimizing a pairwise contrastive (dis)similarity loss at the pixel level between two views of input images. Compared to the baseline method MoCo-v2, our method introduces negligible computation overhead (only <1% slower), but demonstrates consistently superior performance when transferring to downstream dense prediction tasks including object detection, semantic segmentation and instance segmentation; and outperforms the state-of-the-art methods by a large margin. Specifically, over the strong MoCo-v2 baseline, our method achieves significant improvements of 2.0% AP on PASCAL VOC object detection, 1.1% AP on COCO object detection, 0.9% AP on COCO instance segmentation, 3.0% mIoU on PASCAL VOC semantic segmentation and 1.8% mIoU on Cityscapes semantic segmentation. Code is available at: //git.io/AdelaiDet
In the last decade, numerous supervised deep learning approaches requiring large amounts of labeled data have been proposed for visual-inertial odometry (VIO) and depth map estimation. To overcome the data limitation, self-supervised learning has emerged as a promising alternative, exploiting constraints such as geometric and photometric consistency in the scene. In this study, we introduce a novel self-supervised deep learning-based VIO and depth map recovery approach (SelfVIO) using adversarial training and self-adaptive visual-inertial sensor fusion. SelfVIO learns to jointly estimate 6 degrees-of-freedom (6-DoF) ego-motion and a depth map of the scene from unlabeled monocular RGB image sequences and inertial measurement unit (IMU) readings. The proposed approach is able to perform VIO without the need for IMU intrinsic parameters and/or the extrinsic calibration between the IMU and the camera. estimation and single-view depth recovery network. We provide comprehensive quantitative and qualitative evaluations of the proposed framework comparing its performance with state-of-the-art VIO, VO, and visual simultaneous localization and mapping (VSLAM) approaches on the KITTI, EuRoC and Cityscapes datasets. Detailed comparisons prove that SelfVIO outperforms state-of-the-art VIO approaches in terms of pose estimation and depth recovery, making it a promising approach among existing methods in the literature.
Most existing approaches to disfluency detection heavily rely on human-annotated data, which is expensive to obtain in practice. To tackle the training data bottleneck, we investigate methods for combining multiple self-supervised tasks-i.e., supervised tasks where data can be collected without manual labeling. First, we construct large-scale pseudo training data by randomly adding or deleting words from unlabeled news data, and propose two self-supervised pre-training tasks: (i) tagging task to detect the added noisy words. (ii) sentence classification to distinguish original sentences from grammatically-incorrect sentences. We then combine these two tasks to jointly train a network. The pre-trained network is then fine-tuned using human-annotated disfluency detection training data. Experimental results on the commonly used English Switchboard test set show that our approach can achieve competitive performance compared to the previous systems (trained using the full dataset) by using less than 1% (1000 sentences) of the training data. Our method trained on the full dataset significantly outperforms previous methods, reducing the error by 21% on English Switchboard.
This work tackles the problem of semi-supervised learning of image classifiers. Our main insight is that the field of semi-supervised learning can benefit from the quickly advancing field of self-supervised visual representation learning. Unifying these two approaches, we propose the framework of self-supervised semi-supervised learning ($S^4L$) and use it to derive two novel semi-supervised image classification methods. We demonstrate the effectiveness of these methods in comparison to both carefully tuned baselines, and existing semi-supervised learning methods. We then show that $S^4L$ and existing semi-supervised methods can be jointly trained, yielding a new state-of-the-art result on semi-supervised ILSVRC-2012 with 10% of labels.
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