Search

Additional file 1.

Additional file 2.

Additional file 2.

Bimodal palmprint recognition leverages palmprint and palm vein images simultaneously,which achieves high accuracy by multi-model information fusion and has strong anti-falsification property. In the recognition pipeline, the detection of palm and the alignment of region-of-interest (ROI) are two crucial steps for accurate matching. Most existing methods localize palm ROI by keypoint detection algorithms, however the intrinsic difficulties of keypoint detection tasks make the results unsatisfactory. Besides, the ROI alignment and fusion algorithms at image-level are not fully investigaged.To bridge the gap, in this paper, we propose Bimodal Palmprint Fusion Network (BPFNet) which focuses on ROI localization, alignment and bimodal image fusion.BPFNet is an end-to-end framework containing two subnets: The detection network directly regresses the palmprint ROIs based on bounding box prediction and conducts alignment by translation estimation.In the downstream,the bimodal fusion network implements bimodal ROI image fusion leveraging a novel proposed cross-modal selection scheme. To show the effectiveness of BPFNet,we carry out experiments on the large-scale touchless palmprint datasets CUHKSZ-v1 and TongJi and the proposed method achieves state-of-the-art performances., Extended version of ICONIP 2021 paper

Since human-labeled samples are free for the target set, unsupervised person re-identification (Re-ID) has attracted much attention in recent years, by additionally exploiting the source set. However, due to the differences on camera styles, illumination and backgrounds, there exists a large gap between source domain and target domain, introducing a great challenge on cross-domain matching. To tackle this problem, in this paper we propose a novel method named Dual-stream Reciprocal Disentanglement Learning (DRDL), which is quite efficient in learning domain-invariant features. In DRDL, two encoders are first constructed for id-related and id-unrelated feature extractions, which are respectively measured by their associated classifiers. Furthermore, followed by an adversarial learning strategy, both streams reciprocally and positively effect each other, so that the id-related features and id-unrelated features are completely disentangled from a given image, allowing the encoder to be powerful enough to obtain the discriminative but domain-invariant features. In contrast to existing approaches, our proposed method is free from image generation, which not only reduces the computational complexity remarkably, but also removes redundant information from id-related features. Extensive experiments substantiate the superiority of our proposed method compared with the state-of-the-arts. The source code has been released in https://github.com/lhf12278/DRDL., 14 pages,8 figures

About a century ago, Born proposed a possible matter of state, ferroelectric fluid, might exist if the dipole moment is strong enough. The experimental realisation of such states needs magnifying molecular polar nature to macroscopic scales in liquids. Here, we report on the discovery of a novel chiral liquid matter state, dubbed chiral ferronematic, stabilized by the local ferroelectric ordering coupled to the chiral helicity. It carries the polar vector rotating helically, corresponding to a helieletric structure, analogous to the magnetic counterpart of helimagnet. The state can be retained down to room-temperature and demonstrates gigantic dielectric and nonlinear optical responses. The novel matter state opens a new chapter for exploring the material space of the diverse ferroelectric liquids.

While Transformer has achieved remarkable performance in various high-level vision tasks, it is still challenging to exploit the full potential of Transformer in image restoration. The crux lies in the limited depth of applying Transformer in the typical encoder-decoder framework for image restoration, resulting from heavy self-attention computation load and inefficient communications across different depth (scales) of layers. In this paper, we present a deep and effective Transformer-based network for image restoration, termed as U2-Former, which is able to employ Transformer as the core operation to perform image restoration in a deep encoding and decoding space. Specifically, it leverages the nested U-shaped structure to facilitate the interactions across different layers with different scales of feature maps. Furthermore, we optimize the computational efficiency for the basic Transformer block by introducing a feature-filtering mechanism to compress the token representation. Apart from the typical supervision ways for image restoration, our U2-Former also performs contrastive learning in multiple aspects to further decouple the noise component from the background image. Extensive experiments on various image restoration tasks, including reflection removal, rain streak removal and dehazing respectively, demonstrate the effectiveness of the proposed U2-Former.

Although equirectangular projection (ERP) is a convenient form to store omnidirectional images (also known as 360-degree images), it is neither equal-area nor conformal, thus not friendly to subsequent visual communication. In the context of image compression, ERP will over-sample and deform things and stuff near the poles, making it difficult for perceptually optimal bit allocation. In conventional 360-degree image compression, techniques such as region-wise packing and tiled representation are introduced to alleviate the over-sampling problem, achieving limited success. In this paper, we make one of the first attempts to learn deep neural networks for omnidirectional image compression. We first describe parametric pseudocylindrical representation as a generalization of common pseudocylindrical map projections. A computationally tractable greedy method is presented to determine the (sub)-optimal configuration of the pseudocylindrical representation in terms of a novel proxy objective for rate-distortion performance. We then propose pseudocylindrical convolutions for 360-degree image compression. Under reasonable constraints on the parametric representation, the pseudocylindrical convolution can be efficiently implemented by standard convolution with the so-called pseudocylindrical padding. To demonstrate the feasibility of our idea, we implement an end-to-end 360-degree image compression system, consisting of the learned pseudocylindrical representation, an analysis transform, a non-uniform quantizer, a synthesis transform, and an entropy model. Experimental results on $19,790$ omnidirectional images show that our method achieves consistently better rate-distortion performance than the competing methods. Moreover, the visual quality by our method is significantly improved for all images at all bitrates.

In 3D shape recognition, multi-view based methods leverage human's perspective to analyze 3D shapes and have achieved significant outcomes. Most existing research works in deep learning adopt handcrafted networks as backbones due to their high capacity of feature extraction, and also benefit from ImageNet pretraining. However, whether these network architectures are suitable for 3D analysis or not remains unclear. In this paper, we propose a neural architecture search method named Auto-MVCNN which is particularly designed for optimizing architecture in multi-view 3D shape recognition. Auto-MVCNN extends gradient-based frameworks to process multi-view images, by automatically searching the fusion cell to explore intrinsic correlation among view features. Moreover, we develop an end-to-end scheme to enhance retrieval performance through the trade-off parameter search. Extensive experimental results show that the searched architectures significantly outperform manually designed counterparts in various aspects, and our method achieves state-of-the-art performance at the same time.

Super-high-\k{appa} materials that exhibit exceptionally high dielectric permittivity are recognized as potential candidates for a wide range of next-generation photonic and electronic devices. Generally, the high dielectricity for achieving a high-\k{appa} state requires a low symmetry of materials so that most of the discovered high-\k{appa} materials are symmetry-broken crystals. There are scarce reports on fluidic high-\k{appa} dielectrics. Here we demonstrate a rational molecular design, supported by machine-learning analyses, that introduces high polarity to asymmetric molecules, successfully realizing super-high-\k{appa} fluid materials (dielectric permittivity, {\epsilon} > 104) and strong second harmonic generation with macroscopic spontaneous polar ordering. The polar structures are confirmed to be identical for all the synthesized materials. Our experiments and computational calculation reveal the unique orientational structures coupled with the emerging polarity. Furthermore, adopting this strategy to high-molecular-weight systems additionally extends the novel material category from monomer to polar polymer materials, creating polar soft matters with spontaneous symmetry breaking.

At present, continuous extrusion technology is applied in the industrial production of traditional non-ferrous metal profiles, composite profiles, and thin-walled hollow profiles, such as pipes, belts, wires and rows. This technology yields energy savings, high material utilization rates, automatic production, large manufacturing-product lengths, labor savings, and requires small floor space. The present paper introduces the characteristics of continuous extrusion technology and some industrial application examples, and focuses on the development prospects of multi-blank multi-round continuous extrusion technology.

Supercritical carbon dioxide (SCO2) is of great use in miniature power systems. It obtains the characteristics of high density and low viscosity, which makes it possible to build a compact structure for turbomachinery. For a turbine design, an important issue is to figure out the appropriate solidity of the rotor. The objective of this research is to present the aerodynamic performance and provide the design reference for SCO2 and air radial-inflow turbines considering different solidity structures. For the low solidity case of SCO2 turbine, new splitter structures are proposed to improve its performance. The automatic design and simulation process are established by batch modes in MATLAB. The numerical investigation is based on a 3D viscous compressible N-S equation and the actual fluid property of SCO2 and air. The distributions of flow parameters are first presented. Rotor blade load and aerodynamic force are then thoroughly analyzed and the aerodynamic performances of all cases are obtained. The SCO2 turbine has larger power capacity and higher efficiency while the performance of the air turbine is less affected by rotor solidity. For both SCO2 and air, small solidity can cause the unsatisfactory flow condition at the inlet and the shroud section of the rotor, while large solidity results in the aerodynamic loss at the trailing edge of rotor blade and the hub of rotor outlet. A suction side offset splitter can greatly improve the performance of the low solidity SCO2 turbine.

Multi-view subspace learning (MSL) aims to find a low-dimensional subspace of the data obtained from multiple views. Different from single view case, MSL should take both common and specific knowledge among different views into consideration. To enhance the robustness of model, the complexity, non-consistency and similarity of noise in multi-view data should be fully taken into consideration. Most current MSL methods only assume a simple Gaussian or Laplacian distribution for the noise while neglect the complex noise configurations in each view and noise correlations among different views of practical data. To this issue, this work initiates a MSL method by encoding the multi-view-shared and single-view-specific noise knowledge in data. Specifically, we model data noise in each view as a separated Mixture of Gaussians (MoG), which can fit a wider range of complex noise types than conventional Gaussian/Laplacian. Furthermore, we link all single-view-noise as a whole by regularizing them by a common MoG component, encoding the shared noise knowledge among them. Such regularization component can be formulated as a concise KL-divergence regularization term under a MAP framework, leading to good interpretation of our model and simple EM-based solving strategy to the problem. Experimental results substantiate the superiority of our method., 8 pages, 6 figures

Robotics has markedly extended the reach of human beings in sensing, manipulating, and transforming the world around us. One of the most inspiring challenges in science and technology is to extend our capacity with tiny robots towards operations at the micro/nanoscales, a dimension where we can directly interact with fundamental biological building blocks. This thesis is devoted to advancing micro/nanorobotics to extend human being’s capacities in following three themes, with a special focus being given to biomedical applications. The first theme focuses on design and fabrication of bio-inspired un-tethered nanorobots with efficient locomotion, adaptive operation, collective regulation, and eventually biological function towards operation in whole blood. The second theme explores the versatility of functional micro/nanorobots to perform diverse tasks including writing (nanolithography), reading (superresolution imaging), destroying (warfare agents), and repairing (surface cracks), all at the micro/nanoscale. The third theme employs self-propelling microrobot as an active delivery technique that autonomously and precisely transports the therapeutic agents inside live animal’s gastrointestinal tract, improving therapeutic efficacy for bacterial infection treatment. This technique opens the door for micro/nanorobots as an active delivery platform for medical treatment and is promising for a wide range of personalized diagnostic and therapeutic applications. Ultimately, micro/nanorobots has the potential to change the game of science, engineering, and medicine by extending our capacity at the micro/nanoscale. The novel applications presented in this thesis are just a few examples showing the power of micro/nanorobots, with countless more avenues waiting to be explored towards living and active matter, and eventually artificial intelligence and synthetic life at the micro/nanoscale.

A full-polarized facet based scattering model (FPFSM) for investigating the electromagnetic (EM) scattering by two-dimensional electrically large sea surfaces with high efficiency at high microwave bands is proposed. For this method, the scattering field over a large sea facet in a diffuse scattering region is numerically deduced according to the Bragg scattering mechanism. In regard to near specular directions, a novel approach is proposed to calculate the scattered field from a sea surface based on the second order small slope approximation (SSA-II), which saves computer memory considerably and is able to analyze the EM scattering by electrically large sea surfaces. The feasibility of this method in evaluating the radar returns from the sea surface is proved by comparing the normalized radar cross sections (NRCS) and the Doppler spectrum with the SSA-II. Then NRCS results in monostatic and bistatic configurations under different polarization states, scattering angles and wind speeds are analyzed as well as the Doppler spectrum at Ka-band. Numerical results show that the FPFSM is a reliable and efficient method to analyze the full-polarized scattering characteristics from electrically large sea surface within a wide frequency range.

Transient self-destroyed micromotors that autonomously disappear in biological media at controlled rates upon completing their task, without leaving a toxic residue, are presented. The propulsion and degradation characteristics of the self-destroyed Mg/ZnO, Mg/Si, and Zn/Fe Janus micromotors and single-component Zn micromotors are described. The degradation of the Janus micromotors relies on the different corrosion rates of their core-shell components. Inductively coupled plasma optical emission spectrometry measurements are used to probe the time-dependent degradation of the different constituents of the micromotors. The toxicity of the transient micromotors is discussed toward their potential use in biomedical applications. This concept of transient micromotors offers considerable potential for diverse practical applications in the near future.

Radiation belt electrons can exhibit different types of pitch angle distributions in response to various magnetospheric processes. Butterfly distributions, characterized by flux minima at pitch angles around 90°, are broadly observed in both the outer and inner belts and the slot region. Butterfly distributions close to the outer magnetospheric boundary have been attributed to drift shell splitting and losses to the magnetopause. However, their occurrence in the inner belt and the slot region has hitherto not been resolved. By analyzing the particle and wave data collected by the Van Allen Probes during a geomagnetic storm, we combine test particle calculations and Fokker-Planck simulations to reveal that scattering by equatorial magnetosonic waves is a significant cause for the formation of energetic electron butterfly distributions in the inner magnetosphere. Another event shows that a large-amplitude magnetosonic wave in the outer belt can create electron butterfly distributions in just a few minutes.

Effective and accurate diagnosis of Diabetes Mellitus (DM), as well as its early stage Impaired Glucose Regulation (IGR), has attracted much attention recently. Traditional Chinese Medicine (TCM) [3], [5] etc. has proved that tongue, face and sublingual diagnosis as a noninvasive method is a reasonable way for disease detection. However, most previous works only focus on a single modality (tongue, face or sublingual) for diagnosis, although different modalities may provide complementary information for the diagnosis of DM and IGR. In this paper, we propose a novel multi-modal classification method to discriminate between DM (or IGR) and healthy controls. Specially, the tongue, facial and sublingual images are first collected by using a non-invasive capture device. The color, texture and geometry features of these three types of images are then extracted, respectively. Finally, our so-called multi-modal similar and specific learning (MMSSL) approach is proposed to combine features of tongue, face and sublingual, which not only exploits the correlation but also extracts individual components among them. Experimental results on a dataset consisting of 192 Healthy, 198 DM and 114 IGR samples (all samples were obtained from Guangdong Provincial Hospital of Traditional Chinese Medicine) substantiate the effectiveness and superiority of our proposed method for the diagnosis of DM and IGR, compared to the case of using a single modality., 9 pages, 8 figures, 30 conference

Artificial micromotors, operating on locally supplied fuels and performing complex tasks, offer great potential for diverse biomedical applications, including autonomous delivery and release of therapeutic payloads and cell manipulation. Various types of synthetic motors, utilizing different propulsion mechanisms, have been fabricated to operate in biological matrices. However, the performance of these man-made motors has been tested exclusively under in vitro conditions (outside the body); their behavior and functionalities in an in vivo environment (inside the body) remain unknown. Herein, we report an in vivo study of artificial micromotors in a living organism using a mouse model. Such in vivo evaluation examines the distribution, retention, cargo delivery, and acute toxicity profile of synthetic motors in mouse stomach via oral administration. Using zinc-based micromotors as a model, we demonstrate that the acid-driven propulsion in the stomach effectively enhances the binding and retention of the motors as well as of cargo payloads on the stomach wall. The body of the motors gradually dissolves in the gastric acid, autonomously releasing their carried payloads, leaving nothing toxic behind. This work is anticipated to significantly advance the emerging field of nano/micromotors and to open the door to in vivo evaluation and clinical applications of these synthetic motors.