Unsupervised Learning Of Optical Flow Research Articles

Unsupervised learning of optical flow in a multi-frame dynamic environment using temporal dynamic modeling

For visual estimation of optical flow, which is crucial for various vision analyses, unsupervised learning by view synthesis has emerged as a promising alternative to supervised methods because the ground-truth flow is not readily available in many cases. However, unsupervised learning is likely to be unstable when pixel tracking is lost via occlusion and motion blur, or pixel correspondence is impaired by variations in image content and spatial structure over time. Recognizing that dynamic occlusions and object variations usually exhibit a smooth temporal transition in natural settings, we shifted our focus to model unsupervised learning optical flow from multi-frame sequences of such dynamic scenes. Specifically, we simulated various dynamic scenarios and occlusion phenomena based on Markov property, allowing the model to extract motion laws and thus gain performance in dynamic and occluded areas, which diverges from existing methods without considering temporal dynamics. In addition, we introduced a temporal dynamic model based on a well-designed spatial-temporal dual recurrent block, resulting in a lightweight model structure with fast inference speed. Assuming the temporal smoothness of optical flow, we used the prior motions of adjacent frames to supervise the occluded regions more reliably. Experiments on several optical flow benchmarks demonstrated the effectiveness of our method, as the performance is comparable to several state-of-the-art methods with advantages in memory and computational overhead.

InvFlow: Involution and multi-scale interaction for unsupervised learning of optical flow

The convolution neural network is still the main tool for extracting the image features and the motion features for most of the optical flow models. The convolution neural networks cannot model the long-range dependencies, and more details are lost in deeper layers. All the deficiencies in the extracted features affect the estimated flow. Therefore, in this work, we concentrated on optimizing the convolution neural network in both the encoder and decoder parts to improve the image and motion features. To enhance the image features, we utilize the involution to provide rich features and model the long-range dependencies. In addition, we propose a Multi-Scale-Interaction module which utilizes the self-attention to make an interaction between the feature scales to avoid detail loss. Additionally, we propose a Motion-Features-Optimization block that utilizes the deformable convolution to enhance the motion features. Our model achieves the state-of-the-art performance on Sintel and KITTI 2015 benchmarks.

Regularization for Unsupervised Learning of Optical Flow.

Regularization is an important technique for training deep neural networks. In this paper, we propose a novel shared-weight teacher-student strategy and a content-aware regularization (CAR) module. Based on a tiny, learnable, content-aware mask, CAR is randomly applied to some channels in the convolutional layers during training to be able to guide predictions in a shared-weight teacher-student strategy. CAR prevents motion estimation methods in unsupervised learning from co-adaptation. Extensive experiments on optical flow and scene flow estimation show that our method significantly improves on the performance of the original networks and surpasses other popular regularization methods. The method also surpasses all variants with similar architectures and the supervised PWC-Net on MPI-Sintel and on KITTI. Our method shows strong cross-dataset generalization, i.e., our method solely trained on MPI-Sintel outperforms a similarly trained supervised PWC-Net by 27.9% and 32.9% on KITTI, respectively. Our method uses fewer parameters and less computation, and has faster inference times than the original PWC-Net.

MDFlow: Unsupervised Optical Flow Learning by Reliable Mutual Knowledge Distillation

Recent works have shown that optical flow can be learned by deep networks from unlabelled image pairs based on brightness constancy assumption and smoothness prior. Current approaches additionally impose an augmentation regularization term for continual self-supervision, which has been proved to be effective on difficult matching regions. However, this method also amplify the inevitable mismatch in unsupervised setting, blocking the learning process towards optimal solution. To break the dilemma, we propose a novel mutual distillation framework to transfer reliable knowledge back and forth between the teacher and student networks for alternate improvement. Concretely, taking estimation of off-the-shelf unsupervised approach as pseudo labels, our insight locates at defining a confidence selection mechanism to extract relative good matches, and then add diverse data augmentation for distilling adequate and reliable knowledge from teacher to student. Thanks to the decouple nature of our method, we can choose a stronger student architecture for sufficient learning. Finally, better student prediction is adopted to transfer knowledge back to the efficient teacher without additional costs in real deployment. Rather than formulating it as a supervised task, we find that introducing an extra unsupervised term for multi-target learning achieves best final results. Extensive experiments show that our approach, termed MDFlow, achieves state-of-the-art real-time accuracy and generalization ability on challenging benchmarks. Code is available at https://github.com/ltkong218/MDFlow.

Unsupervised Learning of Optical Flow With Non-Occlusion From Geometry

Optical flow estimation is a fundamental problem of computer vision and has many applications in the fields of robot learning and autonomous driving. This paper reveals novel geometric principles of optical flow based on the insight and detailed definition of non-occlusion. Then, two novel loss functions are proposed for the unsupervised learning of optical flow based on the geometric principles of non-occlusion. Specifically, after the occlusion part of the images are masked, the flowing process of pixel coordinate points is carefully considered and geometric constraints are conducted based on the geometric principles of optical flow. First, the optical flow vectors of neighboring pixel coordinate points in the first frame will not intersect during the pixel displacement to the second frame. Secondly, when the cluster containing adjacent four pixel coordinate points in the first frame moves to the second frame, no other pixel coordinate points will flow into the quadrilateral formed by them. According to the two geometrical constraints, the optical flow non-intersection loss and the optical flow non-blocking loss in the non-occlusion regions are proposed. Two loss functions punish the irregular and inexact optical flows in the non-occlusion regions. The experiments on datasets demonstrated that the proposed unsupervised losses of optical flow based on the geometric principles in non-occlusion regions make the estimated optical flow more refined in detail, and improve the performance of unsupervised learning of optical flow. In addition, the experiments training on synthetic data and evaluating on real data show that the generalization ability of optical flow network is improved by our proposed unsupervised approach. Our codes are open sourced at: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/IRMVLab/NccFlow</uri> .

ASFlow: Unsupervised Optical Flow Learning With Adaptive Pyramid Sampling

We present an unsupervised optical flow estimation method by proposing an adaptive pyramid sampling in the deep pyramid network. Specifically, in the pyramid downsampling, we propose a Content-Aware Pooling (CAP) module, which promotes local feature gathering by avoiding cross region pooling, so that the learned features become more representative. In the pyramid upsampling, we propose an Adaptive Flow Upsampling (AFU) module, where cross edge interpolation can be avoided, producing sharp motion boundaries. Equipped with these two modules, our method achieves the best performance for unsupervised optical flow estimation on multiple leading benchmarks, including MPI-Sintel, KITTI 2012 and KITTI 2015. Particularly, we achieve EPE=1.5 on KITTI 2012 and F1=9.67% KITTI 2015, which outperform the previous state-of-the-art methods by 16.7% and 13.1%, respectively.

UnFlowGAN: Unsupervised Learning of Optical Flow with Adversarial Learning

UnFlowGAN: Unsupervised Learning of Optical Flow with Adversarial Learning

Rigid-aware self-supervised GAN for camera ego-motion estimation

Learning-based camera ego-motion estimation has attracted increasing attention and has made impressive improvements. However, the accuracy of unsupervised paradigm is still limited, especially in complex dynamic environment. In this paper, we propose a rigid-aware self-supervised generative adversarial network (GAN) for camera ego-motion estimation, which can effectively learn the rigidity of the scene and improve the accuracy of ego-motion estimation by combining the pixel- and the structure-level perception. Specifically, a rigid-aware generator is first designed for joint unsupervised learning of optical flow, stereo depth and camera pose from two consecutive frames. Then, an iterative pose refinement strategy with rigidity learning is presented to reduce the impact of moving objects in scenes. To overcome the limitation of the purely pixel-wised photometric methods, a rigidity mask embedded discriminator is attached to perceive structural distortion artifacts in synthesized fake images, which encourages the generator to learn additional structure-level information to improve the accuracy of pose estimation. Experiments on the benchmark datasets show that our model achieves the state-of-the-art performance in terms of both RPE and ATE compared to recent GAN-based methods.

OIFlow: Occlusion-Inpainting Optical Flow Estimation by Unsupervised Learning

Occlusion is an inevitable and critical problem in unsupervised optical flow learning. Existing methods either treat occlusions equally as non-occluded regions or simply remove them to avoid incorrectness. However, the occlusion regions can provide effective information for optical flow learning. In this paper, we present OIFlow, an occlusion-inpainting framework to make full use of occlusion regions. Specifically, a new appearance-flow network is proposed to inpaint occluded flows based on the image content. Moreover, a boundary dilated warp is proposed to deal with occlusions caused by displacement beyond the image border. We conduct experiments on multiple leading flow benchmark datasets such as Flying Chairs, KITTI and MPI-Sintel, which demonstrate that the performance is significantly improved by our proposed occlusion handling framework.

Unsupervised Learning of Depth, Optical Flow and Pose With Occlusion From 3D Geometry

In autonomous driving, monocular sequences contain lots of information. Monocular depth estimation, camera ego-motion estimation and optical flow estimation in consecutive frames are high-profile concerns recently. By analyzing tasks above, pixels in the middle frame are modeled into three parts: the rigid region, the non-rigid region, and the occluded region. In joint unsupervised training of depth and pose, we can segment the occluded region explicitly. The occlusion information is used in unsupervised learning of depth, pose and optical flow, as the image reconstructed by depth-pose and optical flow will be invalid in occluded regions. A less-than-mean mask is designed to further exclude the mismatched pixels interfered with by motion or illumination change in the training of depth and pose networks. This method is also used to exclude some trivial mismatched pixels in the training of the optical flow network. Maximum normalization is proposed for depth smoothness term to restrain depth degradation in textureless regions. In the occluded region, as depth and camera motion can provide more reliable motion estimation, they can be used to instruct unsupervised learning of optical flow. Our experiments in KITTI dataset demonstrate that the model based on three regions, full and explicit segmentation of the occlusion region, the rigid region, and the non-rigid region with corresponding unsupervised losses can improve performance on three tasks significantly. The source code is available at: https://github.com/guangmingw/DOPlearning.

Unsupervised learning of optical flow with patch consistency and occlusion estimation

Recent works have shown that deep networks can be trained for optical flow estimation without supervision. Based on the photometric constancy assumption, most of these methods adopt the reconstruction loss as the supervision by point-based backward warping. Inspired by the traditional patch matching based approaches, we propose a patch-based consistency to improve the vanilla unsupervised learning method Ren et al. [1]. Instead of only comparing the corresponding pixel intensity, we locate the correspondence by using the image patches with census transform, which is more robust for the illumination variation and occlusion. Moreover, a novel parallel branch is devised to estimate a soft occlusion mask jointly in an unsupervised way. The mask is adopted to weight our patch-based consistency loss to alleviate the influence of the occlusion. The plenty of experiments have been implemented on Flying Chairs, KITTI and MPI-Sintel benchmarks. The results show that our method is efficient and outperforms the peer unsupervised learning methods that are using the FlowNet-liked network.

STFlow: Self-Taught Optical Flow Estimation Using Pseudo Labels.

The Deep learning of optical flow has been an active area for its empirical success. For the difficulty of obtaining accurate dense correspondence labels, unsupervised learning of optical flow has drawn more and more attention, while the accuracy is still far from satisfaction. By holding the philosophy that better estimation models can be trained with betterapproximated labels, which in turn can be obtained from better estimation models, we propose a self-taught learning framework to continually improve the accuracy using self-generated pseudo labels. The estimated optical flow is first filtered by bidirectional flow consistency validation and occlusion-aware dense labels are then generated by edge-aware interpolation from selected sparse matches. Moreover, by combining reconstruction loss with regression loss on the generated pseudo labels, the performance is further improved. The experimental results demonstrate that our models achieve state-of-the-art results among unsupervised methods on the public KITTI, MPI-Sintel and Flying Chairs datasets.

Unsupervised Learning of Optical Flow With CNN-based Non-Local Filtering.

Estimating optical flow from successive video frames is one of the fundamental problems in computer vision and image processing. In the era of deep learning, many methods have been proposed to use convolutional neural networks (CNNs) for optical flow estimation in an unsupervised manner. However, the performance of unsupervised optical flow approaches is still unsatisfactory and often lagging far behind their supervised counterparts, primarily due to over-smoothing across motion boundaries and occlusion. To address these issues, in this paper, we propose a novel method with a new post-processing term and an effective loss function to estimate optical flow in an unsupervised, end-to-end learning manner. Specifically, we first exploit a CNN-based non-local term to refine the estimated optical flow by removing noise and decreasing blur around motion boundaries. This is implemented via automatically learning weights of dependencies over a large spatial neighborhood. Because of its learning ability, the method is effective for various complicated image sequences. Secondly, to reduce the influence of occlusion, a symmetrical energy formulation is introduced to detect the occlusion map from refined bi-directional optical flows. Then the occlusion map is integrated to the loss function. Extensive experiments are conducted on challenging datasets, i.e. FlyingChairs, MPI-Sintel and KITTI to evaluate the performance of the proposed method. The state-of-the-art results demonstrate the effectiveness of our proposed method.

UnFlow: Unsupervised Learning of Optical Flow With a Bidirectional Census Loss

In the era of end-to-end deep learning, many advances in computer vision are driven by large amounts of labeled data. In the optical flow setting, however, obtaining dense per-pixel ground truth for real scenes is difficult and thus such data is rare. Therefore, recent end-to-end convolutional networks for optical flow rely on synthetic datasets for supervision, but the domain mismatch between training and test scenarios continues to be a challenge. Inspired by classical energy-based optical flow methods, we design an unsupervised loss based on occlusion-aware bidirectional flow estimation and the robust census transform to circumvent the need for ground truth flow. On the KITTI benchmarks, our unsupervised approach outperforms previous unsupervised deep networks by a large margin, and is even more accurate than similar supervised methods trained on synthetic datasets alone. By optionally fine-tuning on the KITTI training data, our method achieves competitive optical flow accuracy on the KITTI 2012 and 2015 benchmarks, thus in addition enabling generic pre-training of supervised networks for datasets with limited amounts of ground truth.