Facial expression recognition is of significant importance in criminal investigation and digital entertainment. Under unconstrained conditions, existing datasets are highly class-imbalanced, the similarity between expressions high. Previous methods tend to improve performance facial through deeper or wider network structures, resulting increased storage computing costs. In this paper, we propose a new adaptive supervised objective named AdaReg loss, re-weighting category coefficients address...

Sketch-photo synthesis plays an important role in sketch-based face photo retrieval and photo-based sketch systems. In this paper, we propose automatic sketch-photo algorithm based on sparse representation. The proposed method works at patch level is composed of two steps: neighbor selection (SNS) for initial estimate the pseudoimage (pseudosketch or pseudophoto) sparse-representation-based enhancement (SRE) further improving quality synthesized image. SNS can find closely related neighbors...

Face sketch-photo synthesis plays a critical role in many applications, such as law enforcement and digital entertainment. Recently, face methods have been proposed under the framework of inductive learning, these obtained promising performance. However, learning-based may result high losses for test samples, because learning minimizes empirical loss training samples. This paper presents novel transductive method that incorporates given samples into process optimizes performance on In...

Exemplar-based face sketch synthesis has been widely applied to both digital entertainment and law enforcement. In this paper, we propose a Bayesian framework for synthesis, which provides systematic interpretation understanding the common properties intrinsic difference in different methods from perspective of probabilistic graphical models. The proposed consists two parts: neighbor selection model weight computation model. Within framework, further method. essential rationale behind method...

Heterogeneous face recognition (HFR) refers to matching images acquired from different sources (i.e., sensors or wavelengths) for identification. HFR plays an important role in both biometrics research and industry. In spite of promising progresses achieved recent years, is still a challenging problem due the difficulty represent two heterogeneous homogeneous manner. Existing methods either image ignoring spatial information, rely on transformation procedure which complicates task....

Face sketch-photo synthesis plays an important role in law enforcement and digital entertainment. Most of the existing methods only use pixel intensities as feature. Since face images can be described using features from multiple aspects, this paper presents a novel representations-based sketch-photo-synthesis method that adaptively combines representations to represent image patch. In particular, it processed filters deploys Markov networks exploit interacting relationships between...

The existence of cancer stem cells (CSCs) in non-small cell lung (NSCLC) has profound implications for therapy. In this study, a disulfiram/copper (DSF/Cu) complex was evaluated vitro and vivo its efficacy to inhibit CSCs, which drive recurrence NSCLC. First, we investigated whether DSF/Cu could ALDH-positive NSCLC tumors derived from sorted CSCs vivo. (0.5/1 μmol/l) significantly inhibited the expression transcription factors (Sox2, Oct-4 Nanog) reduced capacities self-renewal,...

Neural network learning for face sketch synthesis from photos has attracted substantial attention due to its favorable performance. However, most existing deep-learning-based models stacked only by multiple convolutional layers without structured regression often lose the common facial structures, limiting their flexibility in a wide range of practical applications, including intelligent security and digital entertainment. In this article, we introduce neural probabilistic graphical model...

Artistic style transfer aims at migrating the from an example image to a content image. Currently, optimization-based methods have achieved great stylization quality, but expensive time cost restricts their practical applications. Meanwhile, feed-forward still fail synthesize complex style, especially when holistic global and local patterns exist. Inspired by common painting process of drawing draft revising details, we introduce novel method named Laplacian Pyramid Network (LapStyle)....

Person re-identification (Re-ID) aims to retrieve images of the same person across disjoint camera views. Most Re-ID studies focus on pedestrian captured by visible cameras, without considering infrared obtained in dark scenarios. retrieval between and modalities is great significance public security. Current methods usually train a model extract global feature descriptors obtain discriminative representations for (VI-REID). Nevertheless, they ignore detailed information heterogeneous...

Recently, ship detection methods based on deep learning have attracted significant attention due to their superior accuracy over traditional methods. However, there still exist two problems affecting its robustness in practical application. 1) The size of ships one image varies greatly, i.e., different sizes; 2) Numerous gather limited field-of-view, dense distribution. To address these problems, we propose a rotational Libra R-convolutional neural network (CNN) method. Our idea is balance...

With the development of synthetic aperture radar (SAR) imaging and deep learning, SAR ship detection based on convolutional neural networks (CNNs) has been extensively applied in last few years. Nevertheless, there are two main obstacles detection: 1) images have too much noise, such as interference from land area, making it difficult to distinguish objects surrounding background, 2) due multiscale characteristics objects, numerous false negatives results, especially for small objects. To...

In various clinical scenarios, medical image is crucial in disease diagnosis and treatment. Different modalities of images provide complementary information jointly helps doctors to make accurate decision. However, due practical restrictions, certain imaging may be unavailable nor complete. To impute missing data with adequate accuracy, here we propose a framework called self-supervised collaborative learning synthesize modality for images. The proposed method comprehensively utilize all...

Only parts of unlabeled data are selected to train models for most semi-supervised learning methods, whose confidence scores usually higher than the pre-defined threshold (i.e., margin). We argue that recognition performance should be further improved by making full use all data. In this paper, we learn an Adaptive Confidence Margin (Ada-CM) fully leverage deep facial expression recognition. All samples partitioned into two subsets comparing their with adaptively learned margin at each...

Unsupervised person re-identification (Re-Id) has attracted increasing attention due to its practical application in the read-world video surveillance system. The traditional unsupervised Re-Id are mostly based on method alternating between clustering and fine-tuning with classification or metric learning objectives grouped clusters. However, since is an open-set problem, methods often leave out lots of outlier instances group into wrong clusters, thus they can not make full use training...

We present VideoReTalking, a new system to edit the faces of real-world talking head video according input audio, producing high-quality and lip-syncing output even with different emotion. Our disentangles this objective into three sequential tasks: (1) face generation canonical expression; (2) audio-driven lip-sync; (3) enhancement for improving photo-realism. Given talking-head video, we first modify expression each frame same template using editing network, resulting in expression. This...

In label-noise learning, estimating the transition matrix has attracted more and attention as plays an important role in building statistically consistent classifiers. However, it is very challenging to estimate T(x), where x denotes instance, because unidentifiable under instance-dependent noise (IDN). To address this problem, we have noticed that, there are psychological physiological evidences showing that humans likely annotate instances of similar appearances same classes, thus...

With the continuous development of deep learning in field image generation models, a large number vivid forged faces have been generated and spread on Internet. These high-authenticity artifacts could grow into threat to society security. Existing face forgery detection methods directly utilize obtained public shared or centralized data for training but ignore personal privacy security issues when couldn't be centralizedly real-world scenarios. Additionally, different distributions caused by...

Zero-shot learning (ZSL) aims to learn models that can recognize images of semantically related unseen categories, through transferring attribute-based knowledge learned from training data seen classes testing data. As visual attributes play a vital role in ZSL, recent embedding-based methods usually focus on compatibility function between the representation and class semantic attributes. While this work, addition simply region embedding different maintain generalization capability model, we...