A Multifeature Learning and Fusion Network for Facial Age Estimation.

Age estimation from face images has attracted much attention due to its favorable and many real-world applications such as video surveillance and social networking. However, most existing studies usually learn a single kind of age feature and ignore other appearance features such as gender and race, which have a great influence on the age pattern. In this paper, we proposed a compact multifeature learning and fusion method for age estimation. Specifically, we first used three subnetworks to learn gender, race, and age information. Then, we fused these complementary features to further form more robust features for age estimation. Finally, we engineered a regression-ranking age-feature estimator to convert the fusion features into the exact age numbers. Experimental results on three benchmark databases demonstrated the effectiveness and efficiency of the proposed method on facial age estimation in comparison to previous state-of-the-art methods. Moreover, compared with previous state-of-the-art methods, our model was more compact with only a 20 MB memory overhead and is suitable for deployment on mobile or embedded devices for age estimation.

Selection signatures of Fuzhong Buffalo based on whole-genome sequences.

BACKGROUND: Fuzhong buffalo, a native breed of Guangxi Zhuang Autonomous Region, is traditionally used as a draft animal to provide farm power in the rice cultivation. In addition, the Fuzhong buffalo also prepared for the bullfighting festival organized by the locals. The detection of the selective signatures in its genome can help in elucidating the selection mechanisms in its stamina and muscle development of a draft animal. RESULTS: In this study, we analyzed 27 whole genomes of buffalo (including 15 Fuzhong buffalo genomes and 12 published buffalo genomes from Upper Yangtze region). The ZHp, ZFst, π-Ratio, and XP-EHH statistics were used to identify the candidate signatures of positive selection in Fuzhong buffalo. Our results detected a set of candidate genes involving in the pathways and GO terms associated with the response to exercise (e.g., ALDOA, STAT3, AKT2, EIF4E2, CACNA2D2, TCF4, CDH2), immunity (e.g., PTPN22, NKX2-3, PIK3R1, ITK, TMEM173), nervous system (e.g., PTPN21, ROBO1, HOMER1, MAGI2, SLC1A3, NRG3, SNAP47, CTNNA2, ADGRL3). In addition, we also identified several genes related to production and growth traits (e.g., PHLPP1, PRKN, MACF1, UCN3, RALGAPA1, PHKB, PKD1L). Our results depicted several pathways, GO terms, and candidate genes to be associated with response to exercise, immunity, nervous system, and growth traits. CONCLUSIONS: The selective sweep analysis of the Fuzhong buffalo demonstrated positive selection pressure on potential target genes involved in behavior, immunity, and growth traits, etc. Our findings provided a valuable resource for future research on buffalo breeding and an insight into the mechanisms of artificial selection.

Assembly of a miRNA-modified QCM sensor for miRNA recognition through response patterns.

In this paper, a miRNA-based quartz crystal microbalance (QCM) biosensor was fabricated and used to the rapid and effective sensing of miRNA. The specific hybridization between probe miRNA and different selected miRNAs (miR-27a, miR-27b, and Let-7a) cause a different interaction mode, thus display different frequency change and response patterns in the QCM sensor, which were used to detect miR-27a and miR-27b. The selective sensing of miR-27a in mixed miRNA solution was also achieved. This miRNA-based QCM biosensor has the advantages of real-time, label-free, and short cycle detection.

Genome-Wide Detection of Copy Number Variations Associated with Miniature Features in Horses.

Copy number variations (CNVs) are crucial structural genomic variants affecting complex traits in humans and livestock animals. The current study was designed to conduct a comprehensive comparative copy number variation analysis among three breeds, Debao (DB), Baise (BS), and Warmblood (WB), with a specific focus on identifying genomic regions associated with miniature features in horses. Using whole-genome next-generation resequencing data, we identified 18,974 CNVs across 31 autosomes. Among the breeds, we found 4279 breed-specific CNV regions (CNVRs). Baise, Debao, and Warmblood displayed 2978, 986, and 895 distinct CNVRs, respectively, with 202 CNVRs shared across all three breeds. After removing duplicates, we obtained 1545 CNVRs from 26 horse genomes. Functional annotation reveals enrichment in biological functions, including antigen processing, cell metabolism, olfactory conduction, and nervous system development. Debao horses have 970 genes overlapping with CNVRs, possibly causing their small size and mountainous adaptations. We also found that the genes GHR, SOX9, and SOX11 may be responsible for the miniature features of the Debao horse by analyzing their overlapping CNVRs. Overall, this study offers valuable insights into the widespread presence of CNVs in the horse genome. The findings contribute to mapping horse CNVs and advance research on unique miniature traits observed in the Debao horse.

Average Approximate Hashing-Based Double Projections Learning for Cross-Modal Retrieval.

Cross-modal retrieval has attracted considerable attention for searching in large-scale multimedia databases because of its efficiency and effectiveness. As a powerful tool of data analysis, matrix factorization is commonly used to learn hash codes for cross-modal retrieval, but there are still many shortcomings. First, most of these methods only focus on preserving locality of data but they ignore other factors such as preserving reconstruction residual of data during matrix factorization. Second, the energy loss of data is not considered when the data of cross-modal are projected into a common semantic space. Third, the data of cross-modal are directly projected into a unified semantic space which is not reasonable since the data from different modalities have different properties. This article proposes a novel method called average approximate hashing (AAH) to address these problems by: 1) integrating the locality and residual preservation into a graph embedding framework by using the label information; 2) projecting data from different modalities into different semantic spaces and then making the two spaces approximate to each other so that a unified hash code can be obtained; and 3) introducing a principal component analysis (PCA)-like projection matrix into the graph embedding framework to guarantee that the projected data can preserve the main energy of data. AAH obtains the final hash codes by using an average approximate strategy, that is, using the mean of projected data of different modalities as the hash codes. Experiments on standard databases show that the proposed AAH outperforms several state-of-the-art cross-modal hashing methods.

Whole-genome resequencing reveals diversity and selective signals in Longlin goat.

The Longlin goat is one of the most valuable livestock species in Guangxi Autonomous Region of China, but its genomic diversity and selective signals are not clearly elucidated. Here we compared 20 genomes of Longlin goat to 66 genomes of other seven goat breeds worldwide to analyze patterns of Longlin goat genetic variation. We found the lowest linkage disequilibrium at the large distances between SNPs associated with the highest effective population size in the recent generations ago in Longlin goat. The eight goat breeds could be divided into Euro-African and East Asian goat population. Interestingly, like East Asian taurine, the same two migration phases might have occurred in the history of East Asian goat. More importantly, we identified selective signals implicated in immune resistance to disease, especially for skin disease, in Longlin goat. Our findings will not only help understand the evolutionary history and breed characteristic but can provide valuable resources for conservation of germplasm resources and implementation of crossbreeding programs.

Projective Double Reconstructions Based Dictionary Learning Algorithm for Cross-Domain Recognition.

Dictionary learning plays a significant role in the field of machine learning. Existing works mainly focus on learning dictionary from a single domain. In this paper, we propose a novel projective double reconstructions (PDR) based dictionary learning algorithm for cross-domain recognition. Owing the distribution discrepancy between different domains, the label information is hard utilized for improving discriminability of dictionary fully. Thus, we propose a more flexible label consistent term and associate it with each dictionary item, which makes the reconstruction coefficients have more discriminability as much as possible. Due to the intrinsic correlation between cross-domain data, the data should be reconstructed with each other. Based on this consideration, we further propose a projective double reconstructions scheme to guarantee that the learned dictionary has the abilities of data itself reconstruction and data crossreconstruction. This also guarantees that the data from different domains can be boosted mutually for obtaining a good data alignment, making the learned dictionary have more transferability. We integrate the double reconstructions, label consistency constraint and classifier learning into a unified objective and its solution can be obtained by proposed optimization algorithm that is more efficient than the conventional l1 optimization based dictionary learning methods. The experiments show that the proposed PDR not only greatly reduces the time complexity for both training and testing, but also outperforms over the stateof- the-art methods.

Mitogenome Diversity and Maternal Origins of Guangxi Buffalo Breeds.

Guangxi Province, in the southwest of China, is one of the putative migratory corridors or domestication centers for swamp buffalo. In this study, we investigated the evolutionary status of two Guangxi native buffalo breeds (Fuzhong buffalo, n = 15; Xilin buffalo, n = 25) based on the complete mitogenome sequencing. Our results revealed rich genetic diversity in the two buffalo breeds. We detected five haplogroups (SA1, SA2, SB1, SB2, SB3) in the two Guangxi buffalo breeds, and the haplogroup SB3 in the Fuzhong buffalo. Our results showed that the haplogroup SA1 was associated with the major domestication event that involved population expansion in Guangxi buffalo. In conclusion, our findings revealed a high level of maternal genetic diversity and the phylogenetic pattern of the two Guangxi buffalo breeds.

A new QCM signal enhancement strategy based on streptavidin@metal-organic framework complex for miRNA detection.

Sensitive and selective detection of miRNA is of great significance for the early diagnosis of human diseases, especially for cancers. Quartz crystal microbalance (QCM) is an effective tool for detecting biological molecules; however, the application of QCM for miRNA detection is still very limited. One of the great needs for QCM detection is to further improve the QCM signal. Herein, for the first time, we promote a new signal enhancement strategy for the detection of miRNA by QCM. First, a hairpin biotin-modified DNA was used as a probe DNA, which exposes the biotin site when interacting with target miRNA. Then, a streptavidin@metal-organic framework (SA@MOF) complex formed by electrostatic attractions between SA and a MOF was introduced into the QCM detection system. The SA@MOF complexes serve as both a signal amplifier and a specific recognition element via specific biotin-SA interactions. The strategy was applied to the detection of a colorectal cancer marker, miR-221, by using a stable Zr(IV)-MOF, UiO-66-NH2. The detection linear range was 10 fM-1 nM, the detection limit was 6.9 fM, and the relative standard deviation (RSD) (n = 5) was lower than 10% in both simulated conditions and the real serum environment. Furthermore, the detection limit reached 0.79 aM when coupled with the isothermal exponential amplification reaction (EXPAR).

Flexible Affinity Matrix Learning for Unsupervised and Semisupervised Classification.

In this paper, we propose a unified model called flexible affinity matrix learning (FAML) for unsupervised and semisupervised classification by exploiting both the relationship among data and the clustering structure simultaneously. To capture the relationship among data, we exploit the self-expressiveness property of data to learn a structured matrix in which the structures are induced by different norms. A rank constraint is imposed on the Laplacian matrix of the desired affinity matrix, so that the connected components of data are exactly equal to the cluster number. Thus, the clustering structure is explicit in the learned affinity matrix. By making the estimated affinity matrix approximate the structured matrix during the learning procedure, FAML allows the affinity matrix itself to be adaptively adjusted such that the learned affinity matrix can well capture both the relationship among data and the clustering structure. Thus, FAML has the potential to perform better than other related methods. We derive optimization algorithms to solve the corresponding problems. Extensive unsupervised and semisupervised classification experiments on both synthetic data and real-world benchmark data sets show that the proposed FAML consistently outperforms the state-of-the-art methods.

Low-rank and sparse embedding for dimensionality reduction.

In this paper, we propose a robust subspace learning (SL) framework for dimensionality reduction which further extends the existing SL methods to a low-rank and sparse embedding (LRSE) framework from three aspects: overall optimum, robustness and generalization. Owing to the uses of low-rank and sparse constraints, both the global subspaces and local geometric structures of data are captured by the reconstruction coefficient matrix and at the same time the low-dimensional embedding of data are enforced to respect the low-rankness and sparsity. In this way, the reconstruction coefficient matrix learning and SL are jointly performed, which can guarantee an overall optimum. Moreover, we adopt a sparse matrix to model the noise which makes LRSE robust to the different types of noise. The combination of global subspaces and local geometric structures brings better generalization for LRSE than related methods, i.e., LRSE performs better than conventional SL methods in unsupervised and supervised scenarios, particularly in unsupervised scenario the improvement of classification accuracy is considerable. Seven specific SL methods including unsupervised and supervised methods can be derived from the proposed framework and the experiments on different data sets (including corrupted data) demonstrate the superiority of these methods over the existing, well-established SL methods. Further, we exploit experiments to provide some new insights for SL.

Balance Preferences with Performance in Group Role Assignment.

Role assignment is a critical element in the role-based collaboration process. There are many factors to consider when decision makers undertake this task. Such factors include a decision maker's preferences and the team's performance. This paper proposes a series of methods, relative to these factors, to solve the group role assignment with balance problem through an association with the one clause at a time approach that is a well-accepted and logic-based association rule mining method. The proposed methods are verified by simulation experiments. The experimental results present the practicability of the proposed solutions. Using the proposed methods, decision makers need only to establish coarse-grain preferences. The fine-grain preferences can be mined. Furthermore, a balance is obtained between the fine-grain preferences and the team's performance.

Robust Latent Subspace Learning for Image Classification.

This paper proposes a novel method, called robust latent subspace learning (RLSL), for image classification. We formulate an RLSL problem as a joint optimization problem over both the latent SL and classification model parameter predication, which simultaneously minimizes: 1) the regression loss between the learned data representation and objective outputs and 2) the reconstruction error between the learned data representation and original inputs. The latent subspace can be used as a bridge that is expected to seamlessly connect the origin visual features and their class labels and hence improve the overall prediction performance. RLSL combines feature learning with classification so that the learned data representation in the latent subspace is more discriminative for classification. To learn a robust latent subspace, we use a sparse item to compensate error, which helps suppress the interference of noise via weakening its response during regression. An efficient optimization algorithm is designed to solve the proposed optimization problem. To validate the effectiveness of the proposed RLSL method, we conduct experiments on diverse databases and encouraging recognition results are achieved compared with many state-of-the-arts methods.

Orthogonal self-guided similarity preserving projection for classification and clustering.

A suitable feature representation can faithfully preserve the intrinsic structure of data. However, traditional dimensionality reduction (DR) methods commonly use the original input features to define the intrinsic structure, which makes the estimated intrinsic structure unreliable since redundant or noisy features may exist in the original input features. Thus a dilemma is that (1) one needs the most suitable feature representation to define the intrinsic structure of data and (2) one should use the proper intrinsic structure of data to perform feature extraction. To address the problem, in this paper we propose a unified learning framework to simultaneously obtain the optimal feature representation and intrinsic structure of data. The structure is learned from the results of feature learning, and the features are learned to preserve the refined structure of data. By leveraging the interactions between the process of determining the most suitable feature representation and intrinsic structure of data, we can capture accurate structure and obtain the optimal feature representation of data. Experimental results demonstrate that our method outperforms state-of-the-art methods in DR and subspace clustering. The code of the proposed method is available at "http://www.yongxu.org/lunwen.html ".

Data uncertainty in face recognition.

The image of a face varies with the illumination, pose, and facial expression, thus we say that a single face image is of high uncertainty for representing the face. In this sense, a face image is just an observation and it should not be considered as the absolutely accurate representation of the face. As more face images from the same person provide more observations of the face, more face images may be useful for reducing the uncertainty of the representation of the face and improving the accuracy of face recognition. However, in a real world face recognition system, a subject usually has only a limited number of available face images and thus there is high uncertainty. In this paper, we attempt to improve the face recognition accuracy by reducing the uncertainty. First, we reduce the uncertainty of the face representation by synthesizing the virtual training samples. Then, we select useful training samples that are similar to the test sample from the set of all the original and synthesized virtual training samples. Moreover, we state a theorem that determines the upper bound of the number of useful training samples. Finally, we devise a representation approach based on the selected useful training samples to perform face recognition. Experimental results on five widely used face databases demonstrate that our proposed approach can not only obtain a high face recognition accuracy, but also has a lower computational complexity than the other state-of-the-art approaches.