Local features e.g. SIFT and its affine and learned variants provide region-to-region rather than point-to-point correspondences. This has recently been exploited to create new minimal solvers for classical problems such as homography, essential and fundamental matrix estimation. The main advantage of such solvers is that their sample size is smaller, e.g., only two instead of four matches are required to estimate a homography. Works proposing such solvers often claim a significant improvement in run-time thanks to fewer RANSAC iterations. We show that this argument is not valid in practice if the solvers are used naively. To overcome this, we propose guidelines for effective use of region-to-region matches in the course of a full model estimation pipeline. We propose a method for refining the local feature geometries by symmetric intensity-based matching, combine uncertainty propagation inside RANSAC with preemptive model verification, show a general scheme for computing uncertainty of minimal solvers results, and adapt the sample cheirality check for homography estimation. Our experiments show that affine solvers can achieve accuracy comparable to point-based solvers at faster run-times when following our guidelines. We make code available at https://github.com/danini/affine-correspondences-for-camera-geometry.
To be invariant, or not to be invariant: that is the question formulated in this work about local descriptors. A limitation of current feature descriptors is the trade-off between generalization and discriminative power: more invariance means less informative descriptors. We propose to overcome this limitation with a disentanglement of invariance in local descriptors and with an online selection of the most appropriate invariance given the context. Our framework consists in a joint learning of multiple local descriptors with different levels of invariance and of meta descriptors encoding the regional variations of an image. The similarity of these meta descriptors across images is used to select the right invariance when matching the local descriptors. Our approach, named Local Invariance Selection at Runtime for Descriptors (LISRD), enables descriptors to adapt to adverse changes in images, while remaining discriminative when invariance is not required. We demonstrate that our method can boost the performance of current descriptors and outperforms state-of-the-art descriptors in several matching tasks, when evaluated on challenging datasets with day-night illumination as well as viewpoint changes.
We propose a new approach to determine correspondences between image pairs in the wild under large changes in illumination, viewpoint, context, and material. While other approaches find correspondences between pairs of images by treating the images independently, we instead condition on both images to implicitly take account of the differences between them. To achieve this, we introduce (i) a spatial attention mechanism (a co-attention module, CoAM) for conditioning the learned features on both images, and (ii) a distinctiveness score used to choose the best matches at test time. CoAM can be added to standard architectures and trained using self-supervision or supervised data, and achieves a significant performance improvement under hard conditions, e.g. large viewpoint changes. We demonstrate that models using CoAM achieve state of the art or competitive results on a wide range of tasks: local matching, camera localization, 3D reconstruction, and image stylization.
Researchers have attempted utilizing deep neural network (DNN) to learn novel local features from images inspired by its recent successes on a variety of vision tasks. However, existing DNN-based algorithms have not achieved such remarkable progress that could be partly attributed to insufficient utilization of the interactive characters between local feature detector and descriptor. To alleviate these difficulties, we emphasize two desired properties, i.e., repeatability and reliability, to simultaneously summarize the inherent and interactive characters of local feature detector and descriptor. Guided by these properties, a self-supervised framework, namely self-evolving keypoint detection and description (SEKD), is proposed to learn an advanced local feature model from unlabeled natural images. Additionally, to have performance guarantees, novel training strategies have also been dedicatedly designed to minimize the gap between the learned feature and its properties. We benchmark the proposed method on homography estimation, relative pose estimation, and structure-from-motion tasks. Extensive experimental results demonstrate that the proposed method outperforms popular hand-crafted and DNN-based methods by remarkable margins. Ablation studies also verify the effectiveness of each critical training strategy. We will release our code along with the trained model publicly.
Initialization is essential to monocular Simultaneous Localization and Mapping (SLAM) problems. This paper focuses on a novel initialization method for monocular SLAM based on planar features. The algorithm starts by homography estimation in a sliding window. It then proceeds to a global plane optimization (GPO) to obtain camera poses and the plane normal. 3D points can be recovered using planar constraints without triangulation. The proposed method fully exploits the plane information from multiple frames and avoids the ambiguities in homography decomposition. We validate our algorithm on the collected chessboard dataset against baseline implementations and present extensive analysis. Experimental results show that our method outperforms the fine-tuned baselines in both accuracy and real-time.
Modern cameras have limited dynamic ranges and often produce images with saturated or dark regions using a single exposure. Although the problem could be addressed by taking multiple images with different exposures, exposure fusion methods need to deal with ghosting artifacts and detail loss caused by camera motion or moving objects. This paper proposes a deep network for exposure fusion. For reducing the potential ghosting problem, our network only takes two images, an underexposed image and an overexposed one. Our network integrates together homography estimation for compensating camera motion, attention mechanism for correcting remaining misalignment and moving pixels, and adversarial learning for alleviating other remaining artifacts. Experiments on real-world photos taken using handheld mobile phones show that the proposed method can generate high-quality images with faithful detail and vivid color rendition in both dark and bright areas.
Homography estimation is an important step in many computer vision problems. Recently, deep neural network methods have shown to be favorable for this problem when compared to traditional methods. However, these new methods do not consider dynamic content in input images. They train neural networks with only image pairs that can be perfectly aligned using homographies. This paper investigates and discusses how to design and train a deep neural network that handles dynamic scenes. We first collect a large video dataset with dynamic content. We then develop a multi-scale neural network and show that when properly trained using our new dataset, this neural network can already handle dynamic scenes to some extent. To estimate a homography of a dynamic scene in a more principled way, we need to identify the dynamic content. Since dynamic content detection and homography estimation are two tightly coupled tasks, we follow the multi-task learning principles and augment our multi-scale network such that it jointly estimates the dynamics masks and homographies. Our experiments show that our method can robustly estimate homography for challenging scenarios with dynamic scenes, blur artifacts, or lack of textures.
In this paper we attempt to address the problem of geometric multi-model fitting with resorting to a few weakly annotated (WA) data points, which has been sparsely studied so far. In weak annotating, most of the manual annotations are supposed to be correct yet inevitably mixed with incorrect ones. The WA data can be naturally obtained in an interactive way for specific tasks, for example, in the case of homography estimation, one can easily annotate points on the same plane/object with a single label by observing the image. Motivated by this, we propose a novel method to make full use of the WA data to boost the multi-model fitting performance. Specifically, a graph for model proposal sampling is first constructed using the WA data, given the prior that the WA data annotated with the same weak label has a high probability of being assigned to the same model. By incorporating this prior knowledge into the calculation of edge probabilities, vertices (i.e., data points) lie on/near the latent model are likely to connect together and further form a subset/cluster for effective proposals generation. With the proposals generated, the $\alpha$-expansion is adopted for labeling, and our method in return updates the proposals. This works in an iterative way. Extensive experiments validate our method and show that the proposed method produces noticeably better results than state-of-the-art techniques in most cases.
We present a robust estimator for fitting multiple parametric models of the same form to noisy measurements. Applications include finding multiple vanishing points in man-made scenes, fitting planes to architectural imagery, or estimating multiple rigid motions within the same sequence. In contrast to previous works, which resorted to hand-crafted search strategies for multiple model detection, we learn the search strategy from data. A neural network conditioned on previously detected models guides a RANSAC estimator to different subsets of all measurements, thereby finding model instances one after another. We train our method supervised as well as self-supervised. For supervised training of the search strategy, we contribute a new dataset for vanishing point estimation. Leveraging this dataset, the proposed algorithm is superior with respect to other robust estimators as well as to designated vanishing point estimation algorithms. For self-supervised learning of the search, we evaluate the proposed algorithm on multi-homography estimation and demonstrate an accuracy that is superior to state-of-the-art methods.
Identifying salient points in images is a crucial component for visual odometry, Structure-from-Motion or SLAM algorithms. Recently, several learned keypoint methods have demonstrated compelling performance on challenging benchmarks. However, generating consistent and accurate training data for interest-point detection in natural images still remains challenging, especially for human annotators. We introduce IO-Net (i.e. InlierOutlierNet), a novel proxy task for the self-supervision of keypoint detection, description and matching. By making the sampling of inlier-outlier sets from point-pair correspondences fully differentiable within the keypoint learning framework, we show that are able to simultaneously self-supervise keypoint description and improve keypoint matching. Second, we introduce KeyPointNet, a keypoint-network architecture that is especially amenable to robust keypoint detection and description. We design the network to allow local keypoint aggregation to avoid artifacts due to spatial discretizations commonly used for this task, and we improve fine-grained keypoint descriptor performance by taking advantage of efficient sub-pixel convolutions to upsample the descriptor feature-maps to a higher operating resolution. Through extensive experiments and ablative analysis, we show that the proposed self-supervised keypoint learning method greatly improves the quality of feature matching and homography estimation on challenging benchmarks over the state-of-the-art.
Image alignment by mesh warps, such as meshflow, is a fundamental task which has been widely applied in various vision applications(e.g., multi-frame HDR/denoising, video stabilization). Traditional mesh warp methods detect and match image features, where the quality of alignment highly depends on the quality of image features. However, the image features are not robust in occurrence of low-texture and low-light scenes. Deep homography methods, on the other hand, are free from such problem by learning deep features for robust performance. However, a homography is limited to plane motions. In this work, we present a deep meshflow motion model, which takes two images as input and output a sparse motion field with motions located at mesh vertexes. The deep meshflow enjoys the merics of meshflow that can describe nonlinear motions while also shares advantage of deep homography that is robust against challenging textureless scenarios. In particular, a new unsupervised network structure is presented with content-adaptive capability. On one hand, the image content that cannot be aligned under mesh representation are rejected by our learned mask, similar to the RANSAC procedure. On the other hand, we learn multiple mesh resolutions, combining to a non-uniform mesh division. Moreover, a comprehensive dataset is presented, covering various scenes for training and testing. The comparison between both traditional mesh warp methods and deep based methods show the effectiveness of our deep meshflow motion model.
Precise homography estimation between multiple images is a pre-requisite for many computer vision applications. One application that is particularly relevant in today's digital era is the alignment of scanned or camera-captured document images such as insurance claim forms for information extraction. Traditional learning based approaches perform poorly due to the absence of an appropriate gradient. Feature based keypoint extraction techniques for homography estimation in real scene images either detect an extremely large number of inconsistent keypoints due to sharp textual edges, or produce inaccurate keypoint correspondences due to variations in illumination and viewpoint differences between document images. In this paper, we propose a novel algorithm for aligning scanned or camera-captured document images using character based keypoints and a reference template. The algorithm is both fast and accurate and utilizes a standard Optical character recognition (OCR) engine such as Tesseract to find character based unambiguous keypoints, which are utilized to identify precise keypoint correspondences between two images. Finally, the keypoints are used to compute the homography mapping between a test document and a template. We evaluated the proposed approach for information extraction on two real world anonymized datasets comprised of health insurance claim forms and the results support the viability of the proposed technique.
Homography estimation is a basic image alignment method in many applications. It is usually conducted by extracting and matching sparse feature points, which are error-prone in low-light and low-texture images. On the other hand, previous deep homography approaches use either synthetic images for supervised learning or aerial images for unsupervised learning, both ignoring the importance of handling depth disparities and moving objects in real world applications. To overcome these problems, in this work we propose an unsupervised deep homography method with a new architecture design. In the spirit of the RANSAC procedure in traditional methods, we specifically learn an outlier mask to only select reliable regions for homography estimation. We calculate loss with respect to our learned deep features instead of directly comparing image content as did previously. To achieve the unsupervised training, we also formulate a novel triplet loss customized for our network. We verify our method by conducting comprehensive comparisons on a new dataset that covers a wide range of scenes with varying degrees of difficulties for the task. Experimental results reveal that our method outperforms the state-of-the-art including deep solutions and feature-based solutions.
It is hard to create consistent ground truth data for interest points in natural images, since interest points are hard to define clearly and consistently for a human annotator. This makes interest point detectors non-trivial to build. In this work, we introduce an unsupervised deep learning-based interest point detector and descriptor. Using a self-supervised approach, we utilize a siamese network and a novel loss function that enables interest point scores and positions to be learned automatically. The resulting interest point detector and descriptor is UnsuperPoint. We use regression of point positions to 1) make UnsuperPoint end-to-end trainable and 2) to incorporate non-maximum suppression in the model. Unlike most trainable detectors, it requires no generation of pseudo ground truth points, no structure-from-motion-generated representations and the model is learned from only one round of training. Furthermore, we introduce a novel loss function to regularize network predictions to be uniformly distributed. UnsuperPoint runs in real-time with 323 frames per second (fps) at a resolution of $224\times320$ and 90 fps at $480\times640$. It is comparable or better than state-of-the-art performance when measured for speed, repeatability, localization, matching score and homography estimation on the HPatch dataset.
This paper proposes a geometric interpretation of the angles and scales which the orientation- and scale-covariant feature detectors, e.g. SIFT, provide. Two new general constraints are derived on the scales and rotations which can be used in any geometric model estimation tasks. Using these formulas, two new constraints on homography estimation are introduced. Exploiting the derived equations, a solver for estimating the homography from the minimal number of two correspondences is proposed. Also, it is shown how the normalization of the point correspondences affects the rotation and scale parameters, thus achieving numerically stable results. Due to requiring merely two feature pairs, robust estimators, e.g. RANSAC, do significantly fewer iterations than by using the four-point algorithm. When using covariant features, e.g. SIFT, the information about the scale and orientation is given at no cost. The proposed homography estimation method is tested in a synthetic environment and on publicly available real-world datasets.
In this paper, we introduce the STN-Homography model to directly estimate the homography matrix between image pair. Different most CNN-based homography estimation methods which use an alternative 4-point homography parameterization, we use prove that, after coordinate normalization, the variance of elements of coordinate normalized $3\times3$ homography matrix is very small and suitable to be regressed well with CNN. Based on proposed STN-Homography, we use a hierarchical architecture which stacks several STN-Homography models and successively reduce the estimation error. Effectiveness of the proposed method is shown through experiments on MSCOCO dataset, in which it significantly outperforms the state-of-the-art. The average processing time of our hierarchical STN-Homography with 1 stage is only 4.87 ms on the GPU, and the processing time for hierarchical STN-Homography with 3 stages is 17.85 ms. The code will soon be open sourced.
The technique requires the epipolar geometry to be pre-estimated between each image pair. It exploits the constraints which the camera movement implies, in order to apply a closed-form correction to the parameters of the input affinities. Also, it is shown that the rotations and scales obtained by partially affine-covariant detectors, e.g., AKAZE or SIFT, can be completed to be full affine frames by the proposed algorithm. It is validated both in synthetic experiments and on publicly available real-world datasets that the method always improves the output of the evaluated affine-covariant feature detectors. As a by-product, these detectors are compared and the ones obtaining the most accurate affine frames are reported. For demonstrating the applicability, we show that the proposed technique as a pre-processing step improves the accuracy of pose estimation for a camera rig, surface normal and homography estimation.
Environment perception is an important task with great practical value and bird view is an essential part for creating panoramas of surrounding environment. Due to the large gap and severe deformation between the frontal view and bird view, generating a bird view image from a single frontal view is challenging. To tackle this problem, we propose the BridgeGAN, i.e., a novel generative model for bird view synthesis. First, an intermediate view, i.e., homography view, is introduced to bridge the large gap. Next, conditioned on the three views (frontal view, homography view and bird view) in our task, a multi-GAN based model is proposed to learn the challenging cross-view translation. Extensive experiments conducted on a synthetic dataset have demonstrated that the images generated by our model are much better than those generated by existing methods, with more consistent global appearance and sharper details. Ablation studies and discussions show its reliability and robustness in some challenging cases.
Binary local features represent an effective alternative to real-valued descriptors, leading to comparable results for many visual analysis tasks, while being characterized by significantly lower computational complexity and memory requirements. When dealing with large collections, a more compact representation based on global features is often preferred, which can be obtained from local features by means of, e.g., the Bag-of-Visual-Word (BoVW) model. Several applications, including for example visual sensor networks and mobile augmented reality, require visual features to be transmitted over a bandwidth-limited network, thus calling for coding techniques that aim at reducing the required bit budget, while attaining a target level of efficiency. In this paper we investigate a coding scheme tailored to both local and global binary features, which aims at exploiting both spatial and temporal redundancy by means of intra- and inter-frame coding. In this respect, the proposed coding scheme can be conveniently adopted to support the Analyze-Then-Compress (ATC) paradigm. That is, visual features are extracted from the acquired content, encoded at remote nodes, and finally transmitted to a central controller that performs visual analysis. This is in contrast with the traditional approach, in which visual content is acquired at a node, compressed and then sent to a central unit for further processing, according to the Compress-Then-Analyze (CTA) paradigm. In this paper we experimentally compare ATC and CTA by means of rate-efficiency curves in the context of two different visual analysis tasks: homography estimation and content-based retrieval. Our results show that the novel ATC paradigm based on the proposed coding primitives can be competitive with CTA, especially in bandwidth limited scenarios.
This paper introduces a state-of-the-art video representation and applies it to efficient action recognition and detection. We first propose to improve the popular dense trajectory features by explicit camera motion estimation. More specifically, we extract feature point matches between frames using SURF descriptors and dense optical flow. The matches are used to estimate a homography with RANSAC. To improve the robustness of homography estimation, a human detector is employed to remove outlier matches from the human body as human motion is not constrained by the camera. Trajectories consistent with the homography are considered as due to camera motion, and thus removed. We also use the homography to cancel out camera motion from the optical flow. This results in significant improvement on motion-based HOF and MBH descriptors. We further explore the recent Fisher vector as an alternative feature encoding approach to the standard bag-of-words histogram, and consider different ways to include spatial layout information in these encodings. We present a large and varied set of evaluations, considering (i) classification of short basic actions on six datasets, (ii) localization of such actions in feature-length movies, and (iii) large-scale recognition of complex events. We find that our improved trajectory features significantly outperform previous dense trajectories, and that Fisher vectors are superior to bag-of-words encodings for video recognition tasks. In all three tasks, we show substantial improvements over the state-of-the-art results.