CVPR 2019 论文列表

Light field essentially represents the collection of rays in space. The rays captured by multiple light field cameras form subsets of full rays in 3D space and can be transformed to each other. However, most previous approaches model the projection from an arbitrary point in 3D space to corresponding pixel on the sensor. There are few models on describing the ray sampling and transformation among multiple light field cameras. In the paper, we propose a novel ray-space projection model to transform sets of rays captured by multiple light field cameras in term of the Plucker coordinates. We first derive a 6x6 ray-space intrinsic matrix based on multi-projection-center (MPC) model. A homogeneous ray-space projection matrix and a fundamental matrix are then proposed to establish ray-ray correspondences among multiple light fields. Finally, based on the ray-space projection matrix, a novel camera calibration method is proposed to verify the proposed model. A linear constraint and a ray-ray cost function are established for linear initial solution and non-linear optimization respectively. Experimental results on both synthetic and real light field data have verified the effectiveness and robustness of the proposed model.

This paper presents a novel method for rare event detection from an image pair with class-imbalanced datasets. A straightforward approach for event detection tasks is to train a detection network from a large-scale dataset in an end-to-end manner. However, in many applications such as building change detection on satellite images, few positive samples are available for the training. Moreover, an image pair of scenes contains many trivial events, such as in illumination changes or background motions. These many trivial events and the class imbalance problem lead to false alarms for rare event detection. In order to overcome these difficulties, we propose a novel method to learn disentangled representations from only low-cost negative samples. The proposed method disentangles the different aspects in a pair of observations: variant and invariant factors that represent trivial events and image contents, respectively. The effectiveness of the proposed approach is verified by the quantitative evaluations on four change detection datasets, and the qualitative analysis shows that the proposed method can acquire the representations that disentangle rare events from trivial ones.

The objective of deep metric learning (DML) is to learn embeddings that can capture semantic similarity information among data points. Existing pairwise or tripletwise loss functions used in DML are known to suffer from slow convergence due to a large proportion of trivial pairs or triplets as the model improves. To improve this, rankingmotivated structured losses are proposed recently to incorporate multiple examples and exploit the structured information among them. They converge faster and achieve state-of-the-art performance. In this work, we present two limitations of existing ranking-motivated structured losses and propose a novel ranked list loss to solve both of them. First, given a query, only a fraction of data points is incorporated to build the similarity structure. Consequently, some useful examples are ignored and the structure is less informative. To address this, we propose to build a setbased similarity structure by exploiting all instances in the gallery. The samples are split into a positive set and a negative set. Our objective is to make the query closer to the positive set than to the negative set by a margin. Second, previous methods aim to pull positive pairs as close as possible in the embedding space. As a result, the intraclass data distribution might be dropped. In contrast, we propose to learn a hypersphere for each class in order to preserve the similarity structure inside it. Our extensive experiments show that the proposed method achieves state-of-the-art performance on three widely used benchmarks.

This paper presents the first optimal, maximal likelihood, solution to the triangulation problem for radially distorted cameras. The proposed solution to the two-view triangulation problem minimizes the L2-norm of the reprojection error in the distorted image space. We cast the problem as the search for corrected distorted image points, and we use a Lagrange multiplier formulation to impose the epipolar constraint for undistorted points. For the one-parameter division model, this formulation leads to a system of five quartic polynomial equations in five unknowns, which can be exactly solved using the Groebner basis method. While the proposed Groebner basis solution is provably optimal; it is too slow for practical applications. Therefore, we developed a fast iterative solver to this problem. Extensive empirical tests show that the iterative algorithm delivers the optimal solution virtually every time, thus making it an L2-optimal algorithm de facto. It is iterative in nature, yet in practice, it converges in no more than five iterations. We thoroughly evaluate the proposed method on both synthetic and real-world data, and we show the benefits of performing the triangulation in the distorted space in the presence of radial distortion.

Multiple object video object segmentation is a challenging task, specially for the zero-shot case, when no object mask is given at the initial frame and the model has to find the objects to be segmented along the sequence. In our work, we propose a Recurrent network for multiple object Video Object Segmentation (RVOS) that is fully end-to-end trainable. Our model incorporates recurrence on two different domains: (i) the spatial, which allows to discover the different object instances within a frame, and (ii) the temporal, which allows to keep the coherence of the segmented objects along time. We train RVOS for zero-shot video object segmentation and are the first ones to report quantitative results for DAVIS-2017 and YouTube-VOS benchmarks. Further, we adapt RVOS for one-shot video object segmentation by using the masks obtained in previous time steps as inputs to be processed by the recurrent module. Our model reaches comparable results to state-of-the-art techniques in YouTube-VOS benchmark and outperforms all previous video object segmentation methods not using online learning in the DAVIS-2017 benchmark. Moreover, our model achieves faster inference runtimes than previous methods, reaching 44ms/frame on a P100 GPU.

We present a deep learning method for end-to-end monocular 3D object detection and metric shape retrieval. We propose a novel loss formulation by lifting 2D detection, orientation, and scale estimation into 3D space. Instead of optimizing these quantities separately, the 3D instantiation allows to properly measure the metric misalignment of boxes. We experimentally show that our 10D lifting of sparse 2D Regions of Interests (RoIs) achieves great results both for 6D pose and recovery of the textured metric geometry of instances. This further enables 3D synthetic data augmentation via inpainting recovered meshes directly onto the 2D scenes. We evaluate on KITTI3D against other strong monocular methods and demonstrate that our approach doubles the AP on the 3D pose metrics on the official test set, defining the new state of the art.

The ROI (region-of-interest) based pooling method performs pooling operations on the cropped ROI regions for various samples and has shown great success in the object detection methods. It compresses the model size while preserving the localization accuracy, thus it is useful in the visual tracking field. Though being effective, the ROI-based pooling operation is not yet considered in the correlation filter formula. In this paper, we propose a novel ROI pooled correlation filter (RPCF) algorithm for robust visual tracking. Through mathematical derivations, we show that the ROI-based pooling can be equivalently achieved by enforcing additional constraints on the learned filter weights, which makes the ROI-based pooling feasible on the virtual circular samples. Besides, we develop an efficient joint training formula for the proposed correlation filter algorithm, and derive the Fourier solvers for efficient model training. Finally, we evaluate our RPCF tracker on OTB-2013, OTB-2015 and VOT-2017 benchmark datasets. Experimental results show that our tracker performs favourably against other state-of-the-art trackers.

We present RL-GAN-Net, where a reinforcement learning (RL) agent provides fast and robust control of a generative adversarial network (GAN). Our framework is applied to point cloud shape completion that converts noisy, partial point cloud data into a high-fidelity completed shape by controlling the GAN. While a GAN is unstable and hard to train, we circumvent the problem by (1) training the GAN on the latent space representation whose dimension is reduced compared to the raw point cloud input and (2) using an RL agent to find the correct input to the GAN to generate the latent space representation of the shape that best fits the current input of incomplete point cloud. The suggested pipeline robustly completes point cloud with large missing regions. To the best of our knowledge, this is the first attempt to train an RL agent to control the GAN, which effectively learns the highly nonlinear mapping from the input noise of the GAN to the latent space of point cloud. The RL agent replaces the need for complex optimization and consequently makes our technique real time. Additionally, we demonstrate that our pipelines can be used to enhance the classification accuracy of point cloud with missing data.

RGB images differentiate from depth as they carry more details about the color and texture information, which can be utilized as a vital complement to depth for boosting the performance of 3D semantic scene completion (SSC). SSC is composed of 3D shape completion (SC) and semantic scene labeling while most of the existing approaches use depth as the sole input which causes the performance bottleneck. Moreover, the state-of-the-art methods employ 3D CNNs which have cumbersome networks and tremendous parameters. We introduce a light-weight Dimensional Decomposition Residual network (DDR) for 3D dense prediction tasks. The novel factorized convolution layer is effective for reducing the network parameters, and the proposed multi-scale fusion mechanism for depth and color image can improve the completion and segmentation accuracy simultaneously. Our method demonstrates excellent performance on two public datasets. Compared with the latest method SSCNet, we achieve 5.9% gains in SC-IoU and 5.7% gains in SSC-IOU, albeit with only 21% network parameters and 16.6% FLOPs employed compared with that of SSCNet.

This paper proposes a new end-to-end trainable matching network based on receptive field, RF-Net, to compute sparse correspondence between images. Building end-to-end trainable matching framework is desirable and challenging. The very recent approach, LF-Net, successfully embeds the entire feature extraction pipeline into a jointly trainable pipeline, and produces the state-of-the-art matching results. This paper introduces two modifications to the structure of LF-Net. First, we propose to construct receptive feature maps, which lead to more effective keypoint detection. Second, we introduce a general loss function term, neighbor mask, to facilitate training patch selection. This results in improved stability in descriptor training. We trained RF-Net on the open dataset HPatches, and compared it with other methods on multiple benchmark datasets. Experiments show that RF-Net outperforms existing state-of-the-art methods.

Robust principal component analysis (RPCA) has drawn significant attentions due to its powerful capability in recovering low-rank matrices as well as successful appplications in various real world problems. The current state-of-the-art algorithms usually need to solve singular value decomposition of large matrices, which generally has at least a quadratic or even cubic complexity. This drawback has limited the application of RPCA in solving real world problems. To combat this drawback, in this paper we propose a new type of RPCA method, RES-PCA, which is linearly efficient and scalable in both data size and dimension. For comparison purpose, AltProj, an existing scalable approach to RPCA requires the precise knowlwdge of the true rank; otherwise, it may fail to recover low-rank matrices. By contrast, our method works with or without knowing the true rank; even when both methods work, our method is faster. Extensive experiments have been performed and testified to the effectiveness of proposed method quantitatively and in visual quality, which suggests that our method is suitable to be employed as a light-weight, scalable component for RPCA in any application pipelines.

Modern machine learning datasets can have biases for certain representations that are leveraged by algorithms to achieve high performance without learning to solve the underlying task. This problem is referred to as "representation bias". The question of how to reduce the representation biases of a dataset is investigated and a new dataset REPresentAtion bIas Removal (REPAIR) procedure is proposed. This formulates bias minimization as an optimization problem, seeking a weight distribution that penalizes examples easy for a classifier built on a given feature representation. Bias reduction is then equated to maximizing the ratio between the classification loss on the reweighted dataset and the uncertainty of the ground-truth class labels. This is a minimax problem that REPAIR solves by alternatingly updating classifier parameters and dataset resampling weights, using stochastic gradient descent. An experimental set-up is also introduced to measure the bias of any dataset for a given representation, and the impact of this bias on the performance of recognition models. Experiments with synthetic and action recognition data show that dataset REPAIR can significantly reduce representation bias, and lead to improved generalization of models trained on REPAIRed datasets. The tools used for characterizing representation bias, and the proposed dataset REPAIR algorithm, are available at https://github.com/JerryYLi/Dataset-REPAIR/.

Neural Architecture Search (NAS) is an important yet challenging task in network design due to its high computational consumption. To address this issue, we propose the Reinforced Evolutionary Neural Architecture Search (RENAS), which is an evolutionary method with reinforced mutation for NAS. Our method integrates reinforced mutation into an evolution algorithm for neural architecture exploration, in which a mutation controller is introduced to learn the effects of slight modifications and make mutation actions. The reinforced mutation controller guides the model population to evolve efficiently. Furthermore, as child models can inherit parameters from their parents during evolution, our method requires very limited computational resources. In experiments, we conduct the proposed search method on CIFAR-10 and obtain a powerful network architecture, RENASNet. This architecture achieves a competitive result on CIFAR-10. The explored network architecture is transferable to ImageNet and achieves a new state-of-the-art accuracy, i.e., 75.7% top-1 accuracy with 5.36M parameters on mobile ImageNet. We further test its performance on semantic segmentation with DeepLabv3 on the PASCAL VOC. RENASNet outperforms MobileNet-v1, MobileNet-v2 and NASNet. It achieves 75.83% mIOU without being pretrained on COCO.

Dramatic progress has been witnessed in basic vision tasks involving low-level perception, such as object recognition, detection, and tracking. Unfortunately, there is still enormous performance gap between artificial vision systems and human intelligence in terms of higher-level vision problems, especially ones involving reasoning. Earlier attempts in equipping machines with high-level reasoning have hovered around Visual Question Answering (VQA), one typical task associating vision and language understanding. In this work, we propose a new dataset, built in the context of Raven's Progressive Matrices (RPM) and aimed at lifting machine intelligence by associating vision with structural, relational, and analogical reasoning in a hierarchical representation. Unlike previous works in measuring abstract reasoning using RPM, we establish a semantic link between vision and reasoning by providing structure representation. This addition enables a new type of abstract reasoning by jointly operating on the structure representation. Machine reasoning ability using modern computer vision is evaluated in this newly proposed dataset. Additionally, we also provide human performance as a reference. Finally, we show consistent improvement across all models by incorporating a simple neural module that combines visual understanding and structure reasoning.

In this paper, we raise a new problem, namely cross model face recognition (CMFR), which has considerable economic and social significance. The core of this problem is to make features extracted from different models comparable. However, the diversity, mainly caused by different application scenarios, frequent version updating, and all sorts of service platforms, obstructs interaction among different models and poses a great challenge. To solve this problem, from the perspective of Bayesian modelling, we propose R3 Adversarial Network (R3AN) which consists of three paths: Reconstruction, Representation and Regression. We also introduce adversarial learning into the reconstruction path for better performance. Comprehensive experiments on public datasets demonstrate the feasibility of interaction among different models with the proposed framework. When updating the gallery, R3AN conducts the feature transformation nearly 10 times faster than ResNet-101. Meanwhile, the transformed feature distribution is very close to that of target model, and its error rate is incredibly reduced by approximately 75% compared with a naive transformation model. Furthermore, we show that face feature can be deciphered into original face image roughly by the reconstruction path, which may give valuable hints for improving the original face recognition models.

Representing procedure text such as recipe for crossmodal retrieval is inherently a difficult problem, not mentioning to generate image from recipe for visualization. This paper studies a new version of GAN, named Recipe Retrieval Generative Adversarial Network (R2GAN), to explore the feasibility of generating image from procedure text for retrieval problem. The motivation of using GAN is twofold: learning compatible cross-modal features in an adversarial way, and explanation of search results by showing the images generated from recipes. The novelty of R2GAN comes from architecture design, specifically a GAN with one generator and dual discriminators is used, which makes the generation of image from recipe a feasible idea. Furthermore, empowered by the generated images, a two-level ranking loss in both embedding and image spaces are considered. These add-ons not only result in excellent retrieval performance, but also generate close-to-realistic food images useful for explaining ranking of recipes. On recipe1M dataset, R2GAN demonstrates high scalability to data size, outperforms all the existing approaches, and generates images intuitive for human to interpret the search results.

Person search has recently gained attention as the novel task of finding a person, provided as a cropped sample, from a gallery of non-cropped images, whereby several other people are also visible. We believe that i. person detection and re-identification should be pursued in a joint optimization framework and that ii. the person search should leverage the query image extensively (e.g. emphasizing unique query patterns). However, so far, no prior art realizes this. We introduce a novel query-guided end-to-end person search network (QEEPS) to address both aspects. We leverage a most recent joint detector and re-identification work, OIM [37]. We extend this with i. a query-guided Siamese squeeze-and-excitation network (QSSE-Net) that uses global context from both the query and gallery images, ii. a query-guided region proposal network (QRPN) to produce query-relevant proposals, and iii. a query-guided similarity subnetwork (QSimNet), to learn a query-guided re-identification score. QEEPS is the first end-to-end query-guided detection and re-id network. On both the most recent CUHK-SYSU [37] and PRW [46] datasets, we outperform the previous state-of-the-art by a large margin.

We present here a method for computational color constancy in which a deep convolutional neural network is trained to detect achromatic pixels in color images after they have been converted to grayscale. The method does not require any information about the illuminant in the scene and relies on the weak assumption, fulfilled by almost all images available on the web, that training images have been approximately balanced. Because of this requirement we define our method as quasi-unsupervised. After training, unbalanced images can be processed thanks to the preliminary conversion to grayscale of the input to the neural network. The results of an extensive experimentation demonstrate that the proposed method is able to outperform the other unsupervised methods in the state of the art being, at the same time, flexible enough to be supervisedly fine-tuned to reach performance comparable with those of the best supervised methods.

Although deep neural networks are highly effective, their high computational and memory costs severely hinder their applications to portable devices. As a consequence, lowbit quantization, which converts a full-precision neural network into a low-bitwidth integer version, has been an active and promising research topic. Existing methods formulate the low-bit quantization of networks as an approximation or optimization problem. Approximation-based methods confront the gradient mismatch problem, while optimizationbased methods are only suitable for quantizing weights and can introduce high computational cost during the training stage. In this paper, we provide a simple and uniform way for weights and activations quantization by formulating it as a differentiable non-linear function. The quantization function is represented as a linear combination of several Sigmoid functions with learnable biases and scales that could be learned in a lossless and end-to-end manner via continuous relaxation of the steepness of Sigmoid functions. Extensive experiments on image classification and object detection tasks show that our quantization networks outperform state-of-the-art methods. We believe that the proposed method will shed new lights on the interpretation of neural network quantization.