Search

Recent deep learning based approaches have shown promising results for the challenging task of inpainting large missing regions in an image. These methods can generate visually plausible image structures and textures, but often create distorted structures or blurry textures inconsistent with surrounding areas. This is mainly due to ineffectiveness of convolutional neural networks in explicitly borrowing or copying information from distant spatial locations. On the other hand, traditional texture and patch synthesis approaches are particularly suitable when it needs to borrow textures from the surrounding regions. Motivated by these observations, we propose a new deep generative model-based approach which can not only synthesize novel image structures but also explicitly utilize surrounding image features as references during network training to make better predictions. The model is a feed-forward, fully convolutional neural network which can process images with multiple holes at arbitrary locations and with variable sizes during the test time. Experiments on multiple datasets including faces (CelebA, CelebA-HQ), textures (DTD) and natural images (ImageNet, Places2) demonstrate that our proposed approach generates higher-quality inpainting results than existing ones. Code, demo and models are available at: https://github.com/JiahuiYu/generative_inpainting., Accepted in CVPR 2018; add CelebA-HQ results; open sourced; interactive demo available: http://jhyu.me/demo

Existing video prediction methods mainly rely on observing multiple historical frames or focus on predicting the next one-frame. In this work, we study the problem of generating consecutive multiple future frames by observing one single still image only. We formulate the multi-frame prediction task as a multiple time step flow (multi-flow) prediction phase followed by a flow-to-frame synthesis phase. The multi-flow prediction is modeled in a variational probabilistic manner with spatial-temporal relationships learned through 3D convolutions. The flow-to-frame synthesis is modeled as a generative process in order to keep the predicted results lying closer to the manifold shape of real video sequence. Such a two-phase design prevents the model from directly looking at the high-dimensional pixel space of the frame sequence and is demonstrated to be more effective in predicting better and diverse results. Extensive experimental results on videos with different types of motion show that the proposed algorithm performs favorably against existing methods in terms of quality, diversity and human perceptual evaluation.

In this work, we address the challenging video scene parsing problem by developing effective representation learning methods given limited parsing annotations. In particular, we contribute two novel methods that constitute a unified parsing framework. (1) \textbf{Predictive feature learning}} from nearly unlimited unlabeled video data. Different from existing methods learning features from single frame parsing, we learn spatiotemporal discriminative features by enforcing a parsing network to predict future frames and their parsing maps (if available) given only historical frames. In this way, the network can effectively learn to capture video dynamics and temporal context, which are critical clues for video scene parsing, without requiring extra manual annotations. (2) \textbf{Prediction steering parsing}} architecture that effectively adapts the learned spatiotemporal features to scene parsing tasks and provides strong guidance for any off-the-shelf parsing model to achieve better video scene parsing performance. Extensive experiments over two challenging datasets, Cityscapes and Camvid, have demonstrated the effectiveness of our methods by showing significant improvement over well-established baselines., Comment: 15 pages, 7 figures, 5 tables, currently v2

Parsing urban scene images benefits many applications, especially self-driving. Most of the current solutions employ generic image parsing models that treat all scales and locations in the images equally and do not consider the geometry property of car-captured urban scene images. Thus, they suffer from heterogeneous object scales caused by perspective projection of cameras on actual scenes and inevitably encounter parsing failures on distant objects as well as other boundary and recognition errors. In this work, we propose a new FoveaNet model to fully exploit the perspective geometry of scene images and address the common failures of generic parsing models. FoveaNet estimates the perspective geometry of a scene image through a convolutional network which integrates supportive evidence from contextual objects within the image. Based on the perspective geometry information, FoveaNet “undoes” the camera perspective projection — analyzing regions in the space of the actual scene, and thus provides much more reliable parsing results. Furthermore, to effectively address the recognition errors, FoveaNet introduces a new dense CRFs model that takes the perspective geometry as a prior potential. We evaluate FoveaNet on two urban scene parsing datasets, Cityspaces and CamVid, which demonstrates that FoveaNet can outperform all the well-established baselines and provide new state-of-the-art performance.

The success of various applications including robotics, digital content creation, and visualization demand a structured and abstract representation of the 3D world from limited sensor data. Inspired by the nature of human perception of 3D shapes as a collection of simple parts, we explore such an abstract shape representation based on primitives. Given a single depth image of an object, we present 3D-PRNN, a generative recurrent neural network that synthesizes multiple plausible shapes composed of a set of primitives. Our generative model encodes symmetry characteristics of common man-made objects, preserves long-range structural coherence, and describes objects of varying complexity with a compact representation. We also propose a method based on Gaussian Fields to generate a large scale dataset of primitive-based shape representations to train our network. We evaluate our approach on a wide range of examples and show that it outperforms nearest-neighbor based shape retrieval methods and is on-par with voxel-based generative models while using a significantly reduced parameter space., ICCV 2017

The ability to edit materials of objects in images is desirable by many content creators. However, this is an extremely challenging task as it requires to disentangle intrinsic physical properties of an image. We propose an end-to-end network architecture that replicates the forward image formation process to accomplish this task. Specifically, given a single image, the network first predicts intrinsic properties, i.e. shape, illumination, and material, which are then provided to a rendering layer. This layer performs in-network image synthesis, thereby enabling the network to understand the physics behind the image formation process. The proposed rendering layer is fully differentiable, supports both diffuse and specular materials, and thus can be applicable in a variety of problem settings. We demonstrate a rich set of visually plausible material editing examples and provide an extensive comparative study., 14 pages, ICCV 2017

Recent progresses on deep discriminative and generative modeling have shown promising results on texture synthesis. However, existing feed-forward based methods trade off generality for efficiency, which suffer from many issues, such as shortage of generality (i.e., build one network per texture), lack of diversity (i.e., always produce visually identical output) and suboptimality (i.e., generate less satisfying visual effects). In this work, we focus on solving these issues for improved texture synthesis. We propose a deep generative feed-forward network which enables efficient synthesis of multiple textures within one single network and meaningful interpolation between them. Meanwhile, a suite of important techniques are introduced to achieve better convergence and diversity. With extensive experiments, we demonstrate the effectiveness of the proposed model and techniques for synthesizing a large number of textures and show its applications with the stylization., Comment: accepted by CVPR2017

This paper presents the first study on forecasting human dynamics from static images. The problem is to input a single RGB image and generate a sequence of upcoming human body poses in 3D. To address the problem, we propose the 3D Pose Forecasting Network (3D-PFNet). Our 3D-PFNet integrates recent advances on single-image human pose estimation and sequence prediction, and converts the 2D predictions into 3D space. We train our 3D-PFNet using a three-step training strategy to leverage a diverse source of training data, including image and video based human pose datasets and 3D motion capture (MoCap) data. We demonstrate competitive performance of our 3D-PFNet on 2D pose forecasting and 3D pose recovery through quantitative and qualitative results., Comment: Accepted in CVPR 2017

Most previous bounding-box-based segmentation methods assume the bounding box tightly covers the object of interest. However it is common that a rectangle input could be too large or too small. In this paper, we propose a novel segmentation approach that uses a rectangle as a soft constraint by transforming it into an Euclidean distance map. A convolutional encoder-decoder network is trained end-to-end by concatenating images with these distance maps as inputs and predicting the object masks as outputs. Our approach gets accurate segmentation results given sloppy rectangles while being general for both interactive segmentation and instance segmentation. We show our network extends to curve-based input without retraining. We further apply our network to instance-level semantic segmentation and resolve any overlap using a conditional random field. Experiments on benchmark datasets demonstrate the effectiveness of the proposed approaches., Comment: BMVC 2017

In this paper, we propose an effective face completion algorithm using a deep generative model. Different from well-studied background completion, the face completion task is more challenging as it often requires to generate semantically new pixels for the missing key components (e.g., eyes and mouths) that contain large appearance variations. Unlike existing nonparametric algorithms that search for patches to synthesize, our algorithm directly generates contents for missing regions based on a neural network. The model is trained with a combination of a reconstruction loss, two adversarial losses and a semantic parsing loss, which ensures pixel faithfulness and local-global contents consistency. With extensive experimental results, we demonstrate qualitatively and quantitatively that our model is able to deal with a large area of missing pixels in arbitrary shapes and generate realistic face completion results., Comment: Accepted by CVPR 2017

MicroRNAs (miRNAs) modulate various genes associated with brain disorders and circulating miRNAs may therefore serve as biomarkers for these neurological diseases. We previously found that the miRNA miR-221-3p was highly expressed in cerebrospinal fluid and the serum of major depressive disorder (MDD) patients. Here, we examined whether miR-221-3p could be used as a biomarker for depressed mood in perioperative patients. We first examined the relative expression of serum miR-221-3p by real-time quantitative PCR in perioperative patients with different degrees of depressive mood assessed by the Patient Health Questionnaire-9 (PHQ-9) diagnostic form. We found that miR-221-3p expression in the mild depressive mood group (PHQ-9 scores 5–9) was 2.21 fold that of the normal group (PHQ-9 scores 0–4) and the moderate＆severe depressive mood group (PHQ-9 scores ≥ 10) showed miR-221-3p expression levels 3.66 fold that of the normal group. Then the absolute quantification of serum miR-221-3p was obtained using an miRNA standard curve. We found that the amount of serum miR-221-3p was positively correlated with depressed mood; when serum miR-221-3p > 1.7 × 107 copies/μg RNA, all indicated PHQ-9 scores were higher than 6. Subsequently, we found that miR-221-3p could indirectly increase the expression of IFN-α (Interferon alpha) in astrocytes by targeting IRF2 (Interferon Regulatory Factor 2) and that miR-221-3p participated in the anti-neuroinflammatory signaling cascades induced by ketamine and paroxetine via the IRF2/IFN-α pathway. Our results indicate that elevated serum miR-221-3p can be used as a biomarker for depressed mood in perioperative patients and that IFN-α-induced NF-κB activation in astrocytes mediated by miR-221-3p targeting of IRF2 may be one of the potential mechanisms.

We develop a deep learning algorithm for contour detection with a fully convolutional encoder-decoder network. Different from previous low-level edge detection, our algorithm focuses on detecting higher-level object contours. Our network is trained end-to-end on PASCAL VOC with refined ground truth from inaccurate polygon annotations, yielding much higher precision in object contour detection than previous methods. We find that the learned model generalizes well to unseen object classes from the same super-categories on MS COCO and can match state-of-the-art edge detection on BSDS500 with fine-tuning. By combining with the multiscale combinatorial grouping algorithm, our method can generate high-quality segmented object proposals, which significantly advance the state-of-the-art on PASCAL VOC (improving average recall from 0.62 to 0.67) with a relatively small amount of candidates ($\sim$1660 per image)., Accepted by CVPR2016 as spotlight

Interactive object selection is a very important research problem and has many applications. Previous algorithms require substantial user interactions to estimate the foreground and background distributions. In this paper, we present a novel deep learning based algorithm which has a much better understanding of objectness and thus can reduce user interactions to just a few clicks. Our algorithm transforms user provided positive and negative clicks into two Euclidean distance maps which are then concatenated with the RGB channels of images to compose (image, user interactions) pairs. We generate many of such pairs by combining several random sampling strategies to model user click patterns and use them to fine tune deep Fully Convolutional Networks (FCNs). Finally the output probability maps of our FCN 8s model is integrated with graph cut optimization to refine the boundary segments. Our model is trained on the PASCAL segmentation dataset and evaluated on other datasets with different object classes. Experimental results on both seen and unseen objects clearly demonstrate that our algorithm has a good generalization ability and is superior to all existing interactive object selection approaches., Computer Vision and Pattern Recognition

Object segmentation is highly desirable for image understanding and editing. Current interactive tools require a great deal of user effort while automatic methods are usually limited to images of special object categories or with high color contrast. In this paper, we propose a data-driven algorithm that uses examples to break through these limits. As similar objects tend to share similar local shapes, we match query image patches with example images in multiscale to enable local shape transfer. The transferred local shape masks constitute a patch-level segmentation solution space and we thus develop a novel cascade algorithm, PatchCut, for coarse-to-fine object segmentation. In each stage of the cascade, local shape mask candidates are selected to refine the estimated segmentation of the previous stage iteratively with color models. Experimental results on various datasets (Weizmann Horse, Fashionista, Object Discovery and PASCAL) demonstrate the effectiveness and robustness of our algorithm.

This paper formulates face labeling as a conditional random field with unary and pairwise classifiers. We develop a novel multi-objective learning method that optimizes a single unified deep convolutional network with two distinct non-structured loss functions: one encoding the unary label likelihoods and the other encoding the pairwise label dependencies. Moreover, we regularize the network by using a nonparametric prior as new input channels in addition to the RGB image, and show that significant performance improvements can be achieved with a much smaller network size. Experiments on both the LFW and Helen datasets demonstrate state-of-the-art results of the proposed algorithm, and accurate labeling results on challenging images can be obtained by the proposed algorithm for real-world applications.

Appearance-based human re-identification is challenging due to different camera characteristics, varying lighting conditions, pose variations across camera views, etc. Recent studies have revealed that color information plays a critical role on performance. However, two problems remain unclear: (1) how do different color descriptors perform under the same scene in re-identification problem? and (2) how can we combine these descriptors without losing their invariance property and distinctiveness power? In this paper, we propose a novel ensemble model that combines different color descriptors in the decision level through metric learning. Experiments show that the proposed system significantly outperforms state-of-the-art algorithms on two challenging datasets (VIPeR and PRID 450S). We have improved the Rank 1 recognition rate on VIPeR dataset by 8.7%.

In this paper, the dataset of normalized difference vegetation indexes (NDVIs) in the arid region in Xinjiang from 2002 to 2015 as well as the climate data from 52 meteorological stations are utilized and the temporal and spatial variations of NDVI in recent years and their response to temperature and precipitation are analyzed in combination with various methods such as the maximum value analysis, correlation analysis and spatial analysis. It is concluded that in the past 14 years, the annual maximum NDVIs of Xinjiang presented a moderate rising tendency; Under the influence of the global background, the temperature and precipitation also showed different degrees of increase, which showed a significant increase in temperature. The annual maximum NDVI had a significant correlation with the annual precipitation (correlation coefficient: 0.634), but no obvious correlation was identified between the annual maximum NDVI and the annual average temperature (correlation coefficient: 0.279). To this end, regarding to the climatic factors, the influence of precipitation on the vegetation cover is higher than that of temperature.

We present Max-Margin Boltzmann Machines (MMBMs) for object segmentation. MMBMs are essentially a class of Conditional Boltzmann Machines that model the joint distribution of hidden variables and output labels conditioned on input observations. In addition to image-to-label connections, we build direct image-to-hidden connections to facilitate global shape prediction, and thus derive a simple Iterated Conditional Modes algorithm for efficient maximum a posteriori inference. We formulate a max-margin objective function for discriminative training, and analyze the effects of different margin functions on learning. We evaluate MMBMs using three datasets against state-of-the-art methods to demonstrate the strength of the proposed algorithms.

Color naming, which relates colors with color names, can help people with a semantic analysis of images in many computer vision applications. In this paper, we propose a novel salient color names based color descriptor (SCNCD) to describe colors. SCNCD utilizes salient color names to guarantee that a higher probability will be assigned to the color name which is nearer to the color. Based on SCNCD, color distributions over color names in different color spaces are then obtained and fused to generate a feature representation. Moreover, the effect of background information is employed and analyzed for person re-identification. With a simple metric learning method, the proposed approach outperforms the state-of-the-art performance (without user’s feedback optimization) on two challenging datasets (VIPeR and PRID 450S). More importantly, the proposed feature can be obtained very fast if we compute SCNCD of each color in advance.

We present a hybrid parametric and nonparametric algorithm, exemplar cut, for generating class-specific object segmentation hypotheses. For the parametric part, we train a pylon model on a hierarchical region tree as the energy function for segmentation. For the nonparametric part, we match the input image with each exemplar by using regions to obtain a score which augments the energy function from the pylon model. Our method thus generates a set of highly plausible segmentation hypotheses by solving a series of exemplar augmented graph cuts. Experimental results on the Graz and PASCAL datasets show that the proposed algorithm achieves favorable segmentation performance against the state-of-the-art methods in terms of visual quality and accuracy.

This paper presents a deep learning model of building up hierarchical image representation. Each layer of hierarchy consists of three components: sparse coding, saliency pooling and local grouping. With sparse coding we identify distinctive coefficients for representing raw features of each lower layer; saliency pooling helps suppress noise and enhance translation invariance of sparse representation; we group locally pooled sparse codes to form more complex representations. Instead of using hand-crafted descriptors, our model learns an effective image representation directly from images in a unsupervised data-driven manner. We evaluate our algorithm with several benchmark databases of object recognition and analyze the contributions of different components. Experimental results show that our algorithm performs favorably against the state-of-the-art methods.

This paper deals with a new problem in face recognition research, in which the enrollment and query face samples are captured under different lighting conditions. In our case, the enrollment samples are visual light (VIS) images, whereas the query samples are taken under near infrared (NIR) condition. It is very difficult to directly match the face samples captured under these two lighting conditions due to their different visual appearances. In this paper, we propose a novel method for synthesizing VIS images from NIR images based on learning the mappings between images of different spectra (i.e., NIR and VIS). In our approach, we reduce the inter-spectral differences significantly, thus allowing effective matching between faces taken under different imaging conditions. Face recognition experiments clearly show the efficacy of the proposed approach.

Face recognition with partial face images is an important problem in face biometrics. The necessity can arise in not so constrained environments such as in surveillance video, or portal video as provided in Multiple Biometrics Grand Challenge (MBGC). Face alignment with partial face images is a key step toward this challenging problem. In this paper, we present a method for partial face alignment based on scale invariant feature transform (SIFT). We first train a reference model using holistic faces, in which the anchor points and their corresponding descriptor subspaces are learned from initial SIFT keypoints and the relationships between the anchor points are also derived. In the alignment stage, correspondences between the learned holistic face model and an input partial face image are established by matching keypoints of the partial face to the anchor points of the learned face model. Furthermore, shape constraint is used to eliminate outlier correspondences and temporal constraint is explored to find more inliers. Alignment is finally accomplished by solving a similarity transform. Experiments on the MBGC near infrared video sequences show the effectiveness of the proposed method, especially when PCA subspace, shape and temporal constraint are utilized.

Face images captured in different spectral bands, e.g. , in visual (VIS) and near infrared (NIR), are said to be heterogeneous. Although a person's face looks different in heterogeneous images, it should be classified as being from the same individual. In this paper, we present a new method, called face analogy , in the analysis-by-synthesis framework, for heterogeneous face mapping, that is, transforming face images from one type to another, and thereby performing heterogeneous face matching. Experiments show promising results.

Structured support vector machine (SSVM) based methods have demonstrated encouraging performance in recent object tracking benchmarks. However, the complex and expensive optimization limits their deployment in real-world applications. In this paper, we present a simple yet efficient dual linear SSVM (DLSSVM) algorithm to enable fast learning and execution during tracking. By analyzing the dual variables, we propose a primal classifier update formula where the learning step size is computed in closed form. This online learning method significantly improves the robustness of the proposed linear SSVM with lower computational cost. Second, we approximate the intersection kernel for feature representations with an explicit feature map to further improve tracking performance. Finally, we extend the proposed DLSSVM tracker with multi-scale estimation to address the "drift" problem. Experimental results on large benchmark datasets with 50 and 100 video sequences show that the proposed DLSSVM tracking algorithm achieves state-of-the-art performance.