Local Adaptive Gradient Variance Attack for Deep Fake Fingerprint Detection

In recent years,deep learning has been the mainstream technology for fingerprint liveness detection(FLD)tasks because of its remarkable performance.However,recent studies have shown that these deep fake fingerprint detection(DFFD)models are not resistant to attacks by adversarial examples,which are generated by the introduction of subtle perturbations in the fingerprint image,allowing the model to make fake judgments.Most of the existing adversarial example generation methods are based on gradient optimization,which is easy to fall into local optimal,resulting in poor transferability of adversarial attacks.In addition,the perturbation added to the blank area of the fingerprint image is easily perceived by the human eye,leading to poor visual quality.In response to the above challenges,this paper proposes a novel adversarial attack method based on local adaptive gradient variance for DFFD.The ridge texture area within the fingerprint image has been identified and designated as the region for perturbation generation.Subsequently,the images are fed into the targeted white-box model,and the gradient direction is optimized to compute gradient variance.Additionally,an adaptive parameter search method is proposed using stochastic gradient ascent to explore the parameter values during adversarial example generation,aiming to maximize adversarial attack performance.Experimental results on two publicly available fingerprint datasets show that ourmethod achieves higher attack transferability and robustness than existing methods,and the perturbation is harder to perceive.

Coverless Information Hiding Based on the Molecular Structure Images of Material

The traditional information hiding methods embed the secret information by modifying the carrier,which will inevitably leave traces of modification on the carrier.In this way,it is hard to resist the detection of steganalysis algorithm.To address this problem,the concept of coverless information hiding was proposed.Coverless information hiding can effectively resist steganalysis algorithm,since it uses unmodified natural stego-carriers to represent and convey confidential information.However,the state-of-the-arts method has a low hidden capacity,which makes it less appealing.Because the pixel values of different regions of the molecular structure images of material(MSIM)are usually different,this paper proposes a novel coverless information hiding method based on MSIM,which utilizes the average value of sub-image’s pixels to represent the secret information,according to the mapping between pixel value intervals and secret information.In addition,we employ a pseudo-random label sequence that is used to determine the position of sub-images to improve the security of the method.And the histogram of the Bag of words model(BOW)is used to determine the number of subimages in the image that convey secret information.Moreover,to improve the retrieval efficiency,we built a multi-level inverted index structure.Furthermore,the proposed method can also be used for other natural images.Compared with the state-of-the-arts,experimental results and analysis manifest that our method has better performance in anti-steganalysis,security and capacity.

All-carbon positive electrodes for stable aluminium batteries

For addressing the critical problems in current collectors in the aluminium batteries,a variety of carbonbased current collectors,including carbon fiber textiles and three-dimensional(3D)biomass-derivative carbon(BDC)networks,are employed for serving as lightweight non-metal current collectors.The results indicate that all the carbon-based current collectors have electrochemical stability in the acidic electrolyte environments.In the assembled aluminium batteries with all-carbon positive electrodes,thermal annealing process on the carbon-based current collectors has substantially promoted the entire electrochemical energy storage performance.Additionally,both the structure configuration and chemical components of the current collectors have also great impact on the rate capability and cycling stability,implying that the 3D BDC networks are more favorable to offer promoted energy storage capability.Implication of the results from suggests that the carbon-based current collectors and all-carbon positive electrodes are able to deliver more advantages in energy storage behaviors in comparison with the traditional positive electrodes with metal Mo current collectors.Such novel strategy promises a new route for fabricating highperformance positive electrodes for stable advanced aluminium batteries.

A Novel Steganography Scheme Combining Coverless Information Hiding and Steganography

At present,the coverless information hiding has been developed.However,due to the limited mapping relationship between secret information and feature selection,it is challenging to further enhance the hiding capacity of coverless information hiding.At the same time,the steganography algorithm based on object detection only hides secret information in foreground objects,which contribute to the steganography capacity is reduced.Since object recognition contains multiple objects and location,secret information can be mapped to object categories,the relationship of location and so on.Therefore,this paper proposes a new steganography algorithm based on object detection and relationship mapping,which integrates coverless information hiding and steganography.In this method,the coverless information hiding is realized by mapping the object type,color and secret information in object detection method.At the same time,the object detection method is used to find the safe area to hide secret messages.The proposed algorithm can not only improve the steganographic capacity of the two information hiding methods but also make the coverless information hiding more secure and robust.

Instance Retrieval Using Region of Interest Based CNN Features

Recently, image representations derived by convolutional neural networks(CNN) have achieved promising performance for instance retrieval, and they outperformthe traditional hand-crafted image features. However, most of existing CNN-based featuresare proposed to describe the entire images, and thus they are less robust to backgroundclutter. This paper proposes a region of interest (RoI)-based deep convolutionalrepresentation for instance retrieval. It first detects the region of interests (RoIs) from animage, and then extracts a set of RoI-based CNN features from the fully-connected layerof CNN. The proposed RoI-based CNN feature describes the patterns of the detected RoIs,so that the visual matching can be implemented at image region-level to effectively identifytarget objects from cluttered backgrounds. Moreover, we test the performance of theproposed RoI-based CNN feature, when it is extracted from different convolutional layersor fully-connected layers. Also, we compare the performance of RoI-based CNN featurewith those of the state-of-the-art CNN features on two instance retrieval benchmarks.Experimental results show that the proposed RoI-based CNN feature provides superiorperformance than the state-of-the-art CNN features for in-stance retrieval.

A Novel Steganography Algorithm Based on Instance Segmentation

Information hiding tends to hide secret information in image area where is rich texture or high frequency,so as to transmit secret information to the recipient without affecting the visual quality of the image and arousing suspicion.We take advantage of the complexity of the object texture and consider that under certain circumstances,the object texture is more complex than the background of the image,so the foreground object is more suitable for steganography than the background.On the basis of instance segmentation,such as Mask R-CNN,the proposed method hides secret information into each object's region by using the masks of instance segmentation,thus realizing the information hiding of the foreground object without background.This method not only makes it more efficient for the receiver to extract information,but also proves to be more secure and robust by experiments.

Blockchain-Based Decentralized Reputation Management System for Internet of Everything in 6G-Enabled Cybertwin Architecture

Internet of Everything(IoE)has emerged as a promising paradigm for the purpose of connecting and exchanging data among physical objects and humans over the Internet,and it can be widely applied in the fields of industry,transportation,commerce,and education.Recently,the emergence of 6G-enabled cybertwin network architecture provides the technical and theoretical foundation for the realization of IoE paradigm.However,the IoE has three open issues in the 6G-enabled cybertwin architecture,i.e.,data authenticity,data storage and node reliability.To address these issues,we propose a blockchain-based decentralized reputation management system(BC-DRMS)for IoE in 6G-enabled Cybertwin architecture.In the proposed BC-DRMS,the traffic data collected from end nodes is stored on the blockchain and the decentralized file system,i.e.,InterPlanetary File System(IPFS),to resist data tampering,and then the data is further processed by the edge clouds and core clouds to provide services to users.Also,a multi-level reputation evaluation scheme is designed to compute the reputation scores of IoE nodes to prevent malicious node attacks.The experiment results and analysis demonstrate that,compared to the traditional centralized reputation management systems(CRMS),the proposed BC-DRMS cannot only address the issues of data authenticity and storage,but also provides high reliability for IoE in 6G-enabled cybertwin architecture.

Vehicle Matching Based on Similarity Metric Learning

With the development of new media technology,vehicle matching plays a further significant role in video surveillance systems.Recent methods explored the vehicle matching based on the feature extraction.Meanwhile,similarity metric learning also has achieved enormous progress in vehicle matching.But most of these methods are less effective in some realistic scenarios where vehicles usually be captured in different times.To address this cross-domain problem,we propose a cross-domain similarity metric learning method that utilizes theGANto generate vehicle imageswith another domain and propose the two-channel Siamese network to learn a similarity metric from both domains(i.e.,Day pattern or Night pattern)for vehicle matching.To exploit properties and relationships among vehicle datasets,we first apply the domain transformer to translate the domain of vehicle images,and then utilize the two-channel Siamese network to extract features from both domains for better feature similarity learning.Experimental results illustrate that our models achieve improvements over state-of-the-arts.

Study on the Development and Protection of the Historical Figures Resources in China＇s Guangxi Border Areas -- Mrs. WA Cultural Resource Is Taken for Example

Guangxi is in the southwest border of China, and Guilin landscape and ethnic customs as the two ＂trump cards＂ have been widely utilized in the communication between Guaaagxi and the outside world. In fact, Guangxi owns many cultural resources such as humanity history, characteristic diet, and local operas to be further promoted, among which historical figures resources are especially abundant. However, the development and protection of the historical figures resources are inconsistent with the current economic and social development, and feature obviously certain hysteresis. In this paper, by taking Mrs. WA who was a heroine of the Zhuang nationality to resist foreign invasions in the Ming Dynasty, the necessary development road of the historical figures resources is explored based on analyzing the current development and protection of the historical figures resources in China＇ s Guangxi border areas, and then innovative concepts and new ideas are proposed with great efforts.

The Review of Secret Image Sharing

Secret image sharing(SIS)is a significant research topic of image information hiding,which divides the image into multiple shares and dis-tributes them to multiple parties for management and preservation.In order to reconstruct the original image,a subset with predetermined number of shares is needed.And just because it is not necessary to use all of the shares to make a reconstruction,SIS creates a high fault tolerance which breaks the limitations of traditional image protection methods,but at the same time,it causes a reduce of safety.Recently,new technologies,such as deep learning and blockchain,have been applied into SIS to improve its efficiency and security.This paper gives an overall review of SIS,discusses four important approaches for SIS,and makes a comparison analysis among them from the perspectives of pixel expansion,tamper resistance,etc.At the end,this paper indicates the possible research directions of SIS in the future.

A Dynamic YOLO-Based Sequence-Matching Model for Efficient Coverless Image Steganography

Many existing coverless steganography methods establish a mapping relationship between cover images and hidden data.One issue with these methods is that as the steganographic capacity increases,the number of images stored in the database grows exponentially.This makes it challenging to build and manage a large image database.To improve the image library utilization and anti-attack capability of the steganography system,we propose an efficient coverless scheme based on dynamically matched substrings.We utilize You Only Look Once(YOLO)for selecting optimal objects and create a mapping dictionary between these objects and scrambling factors.Using this dictionary,each image is effectively assigned to a specific scrambling factor,which is then used to scramble the receiver’s sequence key.To achieve sufficient steganography capability with a limited image library,all substrings of the scrambled sequences have the potential to hide data.After matching the secret information,the ideal number of stego images will be obtained from the database.According to experimental results,this technology outperforms most previous works in terms of data load,transmission security,and hiding capacity.It can recover an average of 79.85%of secret information under typical geometric attacks,and only approximately 200 random images are needed to achieve a capacity of 19 bits per image.

Speed control strategy for tractor assisted driving based on chassis dynamometer test

Realizing automation of the chassis dynamometer and the unmanned test workshop is an inevitable trend in the development of new tractor products.The accuracy of the speed control of the test tractor directly affects the accuracy of the test loading force.In order to meet the purpose of precise control of the test tractor speed on the chassis dynamometer,a fuzzy PID control strategy was developed according to the working principle of assisted driving.On the basis of traditional PID control,the parameters of fuzzy inference module were added for real-time adjustment to achieve faster response to tractor speed changes and more precise control of tractor speed.The Matlab-Cruise co-simulation platform was established for simulation,and the experiment was verified by the tractor chassis dynamometer using the NEDC working condition and tractor ploughing working condition.The results show that both PID control and fuzzy PID control can achieve tractor speed following accuracy of±0.5 km/h.Fuzzy PID control has higher tractor speed following accuracy,faster response when speed changes,less tractor speed fluctuation,and overall control effect is better than PID control.The research results can provide a reference for the realization of the chassis dynamometer unmanned test workshop.

Differences in pseudogene evolution contributed to the contrasting flavors of turnip and Chiifu, two Brassica rapa subspecies

Pseudogenes are important resources for investigation of genome evolution and genomic diversity because they are nonfunctional but have regulatory effects that influence plant adaptation and diversification.However,few systematic comparative analyses of pseudogenes in closely related species have been conducted.Here,we present a turnip(Brassica rapa ssp.rapa)genome sequence and characterize pseudogenes among diploid Brassica species/subspecies.The results revealed that the number of pseudogenes was greatest in Brassica oleracea(CC genome),followed by B.rapa(AA genome)and then Brassica nigra(BB genome),implying that pseudogene differences emerged after species differentiation.In Brassica AA genomes,pseudogenes were distributed asymmetrically on chromosomes because of numerous chromosomal insertions/rearrangements,which contributed to the diversity among subspecies.Pseudogene differences among subspecies were reflected in the flavor-related glucosinolate(GSL)pathway.Specifically,turnip had the highest content of pungent substances,probably because of expansion of the methylthioalkylmalate synthase-encoding gene family in turnips;these genes were converted into pseudogenes in B.rapa ssp.pekinensis(Chiifu).RNA interference-based silencing of the gene encoding 2-oxoglutarate-dependent dioxygenase 2,which is also associated with flavor and anticancer substances in the GSL pathway,resulted in increased abundance of anticancer compounds and decreased pungency of turnip and Chiifu.These findings revealed that pseudogene differences between turnip and Chiifu influenced the evolution of flavor-associated GSL metabolism-related genes,ultimately resulting in the different flavors of turnip and Chiifu.

Environmental factors drive microbial succession and huangjiu flavor formation during raw wheat qu fermentation

As the core raw material of huangjiu(Chinese rice wine)production and brewing,changes of physichemical factors during the fermentation process significantly affect the succession of the microbial community of raw wheat qu(RWQ),also playing important role in huangjiu quality.In this report we analyzed the community structure and driving force RWQ microorganisms with different processes(manual qu,MAQ;mechanical qu,MEQ),and evaluating the flavor of huangjiu.The production procedure was divided into three stages.Dominant microorganisms were Aspergillus,Alternaria,Epicoccus,Bacillus,Pantoea,Saccharopolyspora and Staphylococcus.In early stage,microbial community of MAQ was significantly correlated with temperature and water content,oxygen in middle stage,while no significant driving force in the last stage.Water content and oxygen affected MEQ community significantly in early stage,while in the final stage,fungal community was affected by both temperature and water content,and bacterial community was significantly correlated with oxygen.Glucoamylase activity reached its maximum at 50℃,while theα-amylase activity decreased with the rise of temperature.The highest amylase activity was obtained with 19%initial water addition.Amylase activity showed positive correlation with the thickness of RWQ.The total amount of organic acids was higher in mechanical huangjiu at the end of pre-fermentation.Contents of n-hexanol,ethyl isovalerate,ethyl hexanoate and guaiacol in manual huangjiu were significantly higher than those in mechanical one,while the content of 4-vinylguaiacol was lower.