3D face reconstruction from a single image still suffers from low accuracy and inability to recover textures in invisible regions. In this paper, we propose a method for generating a 3D portrait with complete texture. The coarse face-and-head model and texture parameters are obtained using 3D Morphable Model fitting. We design an image-geometric inverse renderer that acquires normal, albedo, and light to jointly reconstruct the facial details. Then, we use a texture fusion network to extract the valid texture from rendered faces employing different viewpoints. Specifically, this fused texture recovers the invisible region of the input face, which illustrates the realistic surface of our 3D geometric model. Our approach faithfully reconstructs the original face details, including the profiles and the head region. Extensive experiments are performed to demonstrate that our method outperforms state-of-the-art techniques in various challenging scenarios.

Face texture completion plays a significant role in virtual human research, and the quality of face texture needs to be improved urgently. One of the major obstacles to single-face texture generation is that the generated textures are always incomplete for self-occlusion of the input face, and the other is that pixel details are limited by the illumination. To address this, we propose a method for complete face texture generation based on generative adversarial networks. The face parameters obtained from 3D Morphable Model are processed as conditional vectors in the encoder, and the multivariate Gaussian distribution of the latent code is used in the networks to learn the complete texture features. We established a face texture dataset CFT for training the network. Meanwhile, we show the effectiveness of the proposed approach in qualitative and quantitative experiments. The visual results under different tasks show superior performances compared with the state-of-the-art approaches.

Stereo matching CNNs suffer from performance deteriorate when evaluated under different distributions from training data. Previous domain adaptation/generalization methods are hard to maintain a robust performance in different baselines and usually require difficult adversarial optimization or intricate network structure. To solve this problem, we propose Fourier-based amplitude transform (FAT), mapping the source image to the target style without altering semantic content, which requires no training to perform the domain alignment. Specifically, we leverage the Fourier transform and its inverse to swap the low-frequency amplitude component of the source data with the target data. To effectively map style and relieve the artifacts, we introduce two factors to control the replacing area: the distance of HSV distribution between source and target images; and the difference between the source left image and its warped left image. Experiments testify FAT can significantly bridge domain gaps, making source data distribution closer to target data. Furthermore, when only training on synthetic datasets, FAT can also help different baselines achieve competitive cross-domain generalization capabilities on real datasets.

Most deep learning-based state-of-the-art stereo matching methods significantly depend on large-scale datasets. However, it is implausible to collect sufficient real-world samples with dense and clear ground-truth disparity maps in practice. Although synthetic datasets’ appearance has alleviated the demand for extensive real data, there is a domain shift between synthetic and real sets. To tackle this problem, we propose an individually trained synthetic-to-real domain adaptation (SDA) network that maps synthetic images into the real domain. Specifically, our approach translates the data style from synthetic domain to real domain while maintaining the content and the spatial information. First, edge cues are leveraged to guide domain adaptation in preserving the spatial consistency between input and the generated image. Second, we combine the spatial feature transform (SFT) layer to effectively fuse features from the edge map and the source image. Extensive experiments demonstrate that: 1) when only trained on synthetic data and generalized to real data, our model evidently outperforms many state-of-the-art domain adaptation methods; 2) our translated synthetic datasets (TSD) help to improve the generalization capability of any stereo matching CNNs. Codes and data will be available athttps://github.com/Archaic-Atom/SDA_network.

In the existing fifth-generation (5G) communication system, the remote radio unit (RRU) has a large weight, size and power consumption, and there are also compatibility problems between the low-frequency and the high-frequency millimeter wave frequency band system. Based on the microwave photonics technology, a dual-band radio over fiber (RoF) link is proposed, which can realize the simultaneous transmission and channel switching of the low-frequency (3.5 GHz) and high-frequency (28 GHz). In the low-frequency band, the nonlinear distortion is suppressed, and in the high-frequency millimeter wave band, the power fading caused by dispersion is suppressed.

In this paper, a linearized radio-over-fiber (RoF) link is proposed. The suppression for the third-order intermodulation distortion (IMD3) and periodic power fading can be simultaneously realized in the proposed link. Adjustable carrier-to-sideband ratio (CSR) can also be achieved. Compared with the traditional RoF link, the spurious free dynamic range (SFDR) of the proposed link is improved by 19 dB.

Active Shape Model (ASM) has been successfully applied in the segmentation of Diffusion Tensor Magnetic Resonance Image (DT-MRI, referred to as DTI) of brain. However, due to multiple anatomical structure types, irregular shapes, small gray-scale and large amount of these images, perfect segmentation performance could not be achieved. Especially, it is sensitive to initial values with high computational complexity. In this paper, we introduce the gray information of multiple atlases and the prior information of target shapes into the ASM and propose the Multi-Atlas Active Shape Model (referred to as MA-ASM) approach for DTI segmentation. It was evaluated in a manually labeled database with 7 Region of Interest (ROI)s for each of 20 subjects. In comparison with the state of art method of STAPLE (Simultaneous Truth Performance Level Estimation), the proposed algorithm was closer to the manual segmentation shape by subjective visual effects, and had higher overlap rates and lower error detection rates on quantitative analysis than STAPLE.

Saliency detection is a basic and complex technology in computer vision, which can guide computer to extract key information from image by simulating human visual habit. When image characteristics are unevenly distributed, the accuracy of saliency detection methods will decrease. Unfortunately, this issue is common in natural images and forms a challenge for contrast based methods. We propose an iterative dynamic guided filtering approach to analyze saliency cues. A new and simple kernel function is designed by combining the information of filtering results and input image, which can ensure a good structure transfer from input image to guidance image. The saliency of image pixel is defined based on a novel contrast model using image boundary and center regions. At last, we highlight the result by an exponential function. Experimental results show that the proposed method is superior to the others in terms of detection accuracy and recall rate.

Two key challenges lie in the facial attractiveness computation research: the lack of discriminative face representations, and the scarcity of sufficient and complete training data. Motivated by recent promising work in face recognition using deep neural networks to learn effective features, the first challenge is expected to be addressed from a deep learning point of view. A very deep residual network is utilized to enable automatic learning of hierarchical aesthetics representation. The inspiration to deal with the second challenge comes from the natural representation of the training data, where each training face can be associated with a label (score) distribution given by human raters rather than a single label (average score). This paper, therefore, recasts facial attractiveness computation as a label distribution learning problem. Integrating these two ideas, an end-to-end attractiveness learning framework is established. We also perform feature-level fusion by incorporating the low-level geometric features to further improve the computational performance. Extensive experiments are conducted on a standard benchmark, the SCUT-FBP dataset, where our approach shows significant advantages over the other state-of-the-art work.

Facial attractiveness computation is a challenging task because of the lack of labeled data and discriminative features. In this paper, an end-to-end label distribution learning (LDL) framework with deep convolutional neural network (CNN) and geometric features is proposed to meet these two challenges. Different from the previous work, we recast this task as an LDL problem. Compared with the single label regression, the LDL could improve the generalization ability of our model significantly. In addition, we propose some kinds of geometric features as well as an incremental feature selection method, which could select hundred-dimensional discriminative geometric features from an exhaustive pool of raw features. More importantly, we find these selected geometric features are complementary to CNN features. Extensive experiments are carried out on the SCUT-FBP dataset, where our approach achieves superior performance in comparison to the state-of-the-arts.

We demonstrate a novel method for the automatic modulation classification based on a deep learning autoencoder network, trained by a nonnegativity constraint algorithm. The learning algorithm aims to constrain the negative weights, learns features that amount to a part-based representation of data, and disentangles a more meaningful hidden structure. The performance of this algorithm is tested on the fourth-order cumulants of the modulated signals. The results indicate that the autoencoder with nonnegativity constraint (ANC) improves the sparsity and minimizes the reconstruction error in comparison with the conventional sparse autoencoder. The classification accuracy of an ANC based deep network shows improved accuracy under limited signal length and fading channel.

A video taken under the influence of atmospheric turbulence suffers from serious distortion caused by the variation of optical refractive index. In order to reduce geometric distortion and time-space-varying blur, and recover both coarse structure and fine details, a novel turbulence extraction based approach for recovering a latent image from an atmospheric turbulence degraded imagery sequence is proposed. Firstly, a non-rigid image registration method is applied as a preprocessing to reduce geometric deformation. Secondly, the registered image sequence is decomposed into a low-rank background scene component and a sparse turbulent component via matrix decomposition. Different from other approaches, which intend to remove turbulence directly, we manage to extract information of distortion position from the sparse turbulent component to indicate the sharpest turbulence patches. The selected sharpest turbulence patches are then enhanced and fused to generate an enhanced detail layer. Finally, the output image is generated by fusing the deblurred background scene layer and the enhanced detail layer together. Experiments indicate that our approach is capable of significantly alleviating atmospheric turbulence blur and geometric distortion.

Capturing a professional like photo is always a challenging task, especially for novice users. This paper proposes a photography suggestion approach to assist users to take high visual quality photos with human in the scene. In this research, we first investigate a set of aesthetic composition rules and visual perception principles to construct an aesthetic score prediction model in order to measure visual quality in terms of photo composition. Then we conduct a study on professional photos to define a proper size for the enclosure of human into the picture of a given scene. The proposed approach is able to leverage saliency and geometric detection to represent composition features. Finally, we utilize an efficient hierarchical search to obtain the optimal enclosure for human in the scene. Extensive experiments have been performed for hot spot landmark locations. Through subjective evaluation, the results show that the proposed approach can effectively provide appealing composition recommendation to help users take high quality photos with human in the scene.

To pose like a fashion model and to take professional grade photo are always two challenging tasks, especially for novice users. Pose Maker, an innovative system for pose and photo composition recommendation and synthesis, is developed in this work. Given a user-provided clothing color and gender, this system shall not only offer some suitable poses, but also assist users to take high visual quality photos by generating the visual effect of person in the landscape pictures. To recommend poses, we first propose a view specific professional learning model to help users select several compatible poses for a given image. Based on selection results, a hierarchical pose image synthesis module is designed to synthesize the selected poses along with the scene to be captured in the most suitable position and size taking into consideration several important factors in picture composition, including aesthetic photography principles, color harmony and focal length parameters. Extensive experimental evaluations and analysis on test images of various conditions demonstrate the effectiveness of the proposed system.

Hue component is generally denoted by angle value, so the conventional edge detection operators are incapable to accurately detect edges of hue component. As a result, the popular methods of colour image edge detection usually omit the role of hue component, thus missing some edges caused by hue changes. The authors propose a novel colour edge detection method based on the fusion of hue component and principal component analysis to solve the above problems. First, a novel computational method of hue difference is defined, and then it is applied to classical gradient operators to obtain accurate edges for hue component. Moreover, complete object edges can be obtained by using the edge fusion of the first principal component and hue component of colour image with low‐computational complexity. Experimental results show that the proposed approach not only can act on hue component directly and obtain accurate edges caused by hue changes, but also is effective and easy to implement.

The existing lexicographical ordering approaches respect the total ordering properties, thus making this approach a very robust solution for multivariate ordering. However, different marginal components derived from various representations of a colour image will lead to different results of multivariate ordering. Moreover, the output of lexicographical ordering only depends on the first component leading to the followed components taking no effect. To address these issues, three new marginal components are obtained by means of quaternion decomposition, and they are employed by fuzzy lexicographical ordering, and thus a new fuzzy extremum estimation algorithm (FEEA) based on quaternion decomposition is proposed in this study. The novel multivariate mathematical morphological operators are also defined according to FEEA. Comparing with the existing solutions, experimental results show that the proposed FEEA performs better results on multivariate extremum estimation, and the presented multivariate mathematical operators can be easily handled and can provide better results on multivariate image filtering.

This paper presents a novel active contour model for simultaneous segmentation and bias field estimation of medical images. Based on the additive model of images with intensity in homogeneity, we characterize the statistics of image intensities belonging to each different object in local regions as Gaussian distributions with different means and variances. According to maximum a posteriori probability (MAP) and Bayes rule, we first derive a local objective function for image intensities in a neighborhood around each pixel. Then this local objective function is integrated with respect to the neighborhood center over the entire domain to give a global criterion. In a level set formulation, this global criterion defines an energy in terms of the level set functions that represent a partition of the image domain and a bias field that accounts for the intensity in homogeneity of the image. Therefore, image segmentation and bias field estimation are simultaneously achieved via a level set evolution process. Experimental results for synthetic and real images show desirable performances of our method.

Two fundamental problems in vector extremum estimation based on lexicographical ordering are investigated in this paper. The first problem is the choice of the optimal marginal components applying to lexicographical ordering. And the second is that the lexicographical ordering only depends on the first component which leads to the followed components taking no effect. To solve the two problems, three novel marginal components are presented by means of the quaternion description of a color image. Moreover, a new fuzzy extremum estimation algorithm (FEEA) is also proposed and applied to vector ordering. Comparing with the existing solutions, experimental results show that the proposed FEEA has better effect on vector extremum estimation, and the presented frame of vector mathematical morphology can be used to process color image and provide better result.

3D face scans have been widely used for face modeling and face analysis. Due to the fact that face scans provide variable point clouds across frames, they may not capture complete facial data or miss point-to-point correspondences across various facial scans, thus causing difficulties to use such data for analysis. This paper presents an efficient approach to represent facial shapes from face scans through the reconstruction of face models based on regional information and a generic model. A hybrid approach using two vertex mapping algorithms, displacement mapping and point-to-surface mapping, and a regional blending algorithm are proposed to reconstruct the facial surface detail. The resulting models can represent individual facial shapes consistently and adaptively, establishing the facial point correspondence across individual models. The accuracy of the generated models is evaluated quantitatively. The applicability of the models is validated through the application for 3D facial expression recognition based on the databases of static 3DFE and dynamic 4DFE. A comparison with the state of the art has also been reported.

This paper presents visible light positioning (VLP) techniques based on a traffic light and photodiodes. The position of a vehicle can be determined using both the traffic light position information obtained through a visible light communication (VLC) link and the time difference of arrival (TDOA) of the traffic light signal to two photodiodes mounted in the front of the vehicle. Methods for VLP with one traffic light and two traffic lights are both proposed, and the error due to the non-coplanar effect is studied and the related coplanar rotation methods for both cases are developed. The numerical result shows bias performance of the positioning methods with coplanar rotation is much better than the ones without coplanar rotation, and the effects of the vehicle distance to traffic light and vehicle speed are also discussed.

Based on the ideas of feature fusion and Kernel Canonical Correlation Analysis (KCCA), a novel framework for fusing global and local features on Automatic Target Recognition (ATR) algorithm is proposed. Firstly, the feature fusion method based on KCCA is established, then pseudo Zernike moments and Scale Invariant Feature Transform (SIFT) are extracted as global features and local features. K-means algorithm is applied to normalize the local features to obtain the same form as global features. After the fusion of two features, one-against-all Support Vector Machine (SVM) is employed as classifier for the Multi-class target recognition. Theoretical analysis and experiments on aircraft images results show that KCCA features fusion representations significantly outperform CCA fusion method and single feature approach. Feature fusion of global features and local features based on target image for recognition are proved to be a promising strategy in object recognition field.