Search

The NOD-Like Receptor Protein-3 (NLRP3) inflammasome is a key therapeutic target for the treatment of epilepsy and has been reported to regulate inflammation in several neurological diseases. In this study, a machine learning-based virtual screening strategy has investigated candidate active compounds that inhibit the NLRP3 inflammasome. As machine learning-based virtual screening has the potential to accurately predict protein-ligand binding and reduce false positives outcomes compared to traditional virtual screening. Briefly, classification models were created using Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbor (KNN) machine learning methods. To determine the most crucial features of a molecule's activity, feature selection was carried out. By utilizing 10-fold cross-validation, the created models were analyzed. Among the generated models, the RF model obtained the best results as compared to others. Therefore, the RF model was used as a screening tool against the large chemical databases. Molecular operating environment (MOE) and PyRx software's were applied for molecular docking. Also, using the Amber Tools program, molecular dynamics (MD) simulation of potent inhibitors was carried out. The results showed that the KNN, SVM, and RF accuracy was 0.911 %, 0.906 %, and 0.946 %, respectively. Moreover, the model has shown sensitivity of 0.82 %, 0.78 %, and 0.86 % and specificity of 0.95 %, 0.96 %, and 0.98 % respectively. By applying the model to the ZINC and South African databases, we identified 98 and 39 compounds, respectively, potentially possessing anti-NLRP3 activity. Also, a molecular docking analysis produced ten ZINC and seven South African compounds that has comparable binding affinities to the reference drug. Moreover, MD analysis of the two complexes revealed that the two compounds (ZINC000009601348 and SANC00225) form stable complexes with varying amounts of binding energy. The in-silico studies indicate that both compounds most likely display their inhibitory effect by inhibiting the NLRP3 protein.

Abstract Forecasting the subnational population accurately is needed for sustainable development, including planning for the future, allocating resources, or providing health services. Two approaches are used for forecasting subnational populations: local forecasting where a model is trained for each area, and global forecasting, where one model is trained with all areas. Local forecasting (e.g., statistical models) is limited to capturing the population growth patterns in a single area. Machine learning models, such as the light gradient boosting model (LGBM), are considered a more suitable approach for global forecasting, but it is limited to one-step predictions, leading to error accumulation. Also, combining several models into one ensemble model are used which helped in reduce forecasting errors. However, the nature of population growth is nonlinear, and there is a need to reduce error accumulation. This study overcomes these issues and proposes a population fusion transformer (PFT) as a global forecasting model for population forecasting, which outputs multi-step predictions. The PFT is based on a temporal fusion transformer (TFT) proposing a novel deep gated residual network (DGRN) block to capture data nonlinearity. This study also incorporates the proposed PFT model into various ensemble models to reduce forecasting errors using different prediction and learning approaches. The proposed models are applied to four subnational population datasets from several countries. The PFT model outperforms the LGBM and TFT with lower forecasting errors in three and two datasets. More importantly, combining the PFT with other models in ensemble models reduced errors further.

Abstract The signal transducer and activator of transcription 3 (STAT3) plays a fundamental role in the growth and regulation of cellular life. Activation and over-expression of STAT3 have been implicated in many cancers including solid blood tumors and other diseases such as liver fibrosis and rheumatoid arthritis. Therefore, STAT3 inhibitors are be coming a growing and interesting area of pharmacological research. Consequently, the aim of this study is to design novel inhibitors of STAT3-SH3 computationally for the reduction of liver fibrosis. Herein, we performed Pharmacophore-based virtual screening of databases including more than 19,481 commercially available compounds and in-house compounds. The hits obtained from virtual screening were further docked with the STAT3 receptor. The hits were further ranked on the basis of docking score and binding interaction with the active site of STAT3. ADMET properties of the screened compounds were calculated and filtered based on drug-likeness criteria. Finally, the top five drug-like hit compounds were selected and subjected to molecular dynamic simulation. The stability of each drug-like hit in complex with STAT3 was determined by computing their RMSD, RMSF, Rg, and DCCM analyses. Among all the compounds Sa32 revealed a good docking score, interactions, and stability during the entire simulation procedure. As compared to the Reference compound, the drug-like hit compound Sa32 showed good docking scores, interaction, stability, and binding energy. Therefore, we identified Sa32 as the best small molecule potent inhibitor for STAT3 that will be helpful in the future for the treatment of liver fibrosis.

Genetic skeletal disorders (GSDs) are a large group of rare heterogeneous disorders characterized by abnormal development, remodeling, and growth of the human skeleton's cartilage and bones. GSDs have a high spectrum of phenotypes that range from disproportionate short stature (dwarfism) in childhood to osteoarthritis in old age. According to the latest nosology classification of skeletal dysplasias, 461 disorders under 42 groups are classified according to specific radiographic, clinical, and molecular standards. In addition, correct molecular diagnosis for these rare GSDs is important for genetic and psychological counseling and treatment. GSDs are also associated with many syndromic forms that affect other parts such as hearing, vision, neurological, pulmonary, renal, or cardiac function. This review highlights the importance of GSDs and details a few selected disorders and their management strategies. [JBCGenetics 2023; 6(1.000): 57-69]

Background: Glycosylphosphatidylinositol (GPI) is a glycolipid containing phosphatidylinositol related to the protein surfaces by covalent attachment. Inherited GPI deficiencies have various phenotypic chrematistics, which range from intellectual disability to dysmorphic features, epilepsy, and other severe anomalies. Methods: Molecular diagnosis was performed using whole exome sequencing (WES) followed by Sanger sequencing. Results: WES revealed a novel homozygous nonsense variant (c.250C>T; p.Gln84Ter) in the exon 2 of the phosphatidylinositol glycan anchor biosynthesis class Ogene that might explain the disease phenotype in the patient. Conclusion: This study will help in proper genetic counselling of the family and help in genotype-phenotype correlation in the future. [JBCGenetics 2023; 6(1.000): 22-28]

Background: Polydactyly or hexadactyly is a familiar limb defect that either occurs as an isolated entity (non-syndromic) or is associated with severe (syndromic) morphological phenotypes. Generally, it appears due to a defect in the anteroposterior patterning during limb development. Methods: Here, we present a proband having non-syndromic post-axial polydactyly (PAP) evaluated using whole exome sequencing followed by Sanger sequencing. Furthermore, 3D protein modeling was executed for the normal and mutated IQ domain-containing protein E (IQCE) gene. Results: WES analysis revealed an already reported bi-allelic variant (c.395-1 G>A) in the IQCE gene, previously associated with PAP 7. Furthermore, 3D modeling revealed significant fluctuations in the IQCE protein secondary structure, thus affecting downstream signaling. Conclusion: The work presented validated the significant role of the IQCE gene in the development and patterning of human limbs. [JBCGenetics 2023; 6(1.000): 29-35]

Background: Neurodevelopmental disorders are characterized by different combinations of intellectual disability (ID), communication and social skills deficits, and delays in achieving motor or language milestones. SLITRK2 is a postsynaptic cell-adhesion molecule that promotes neurite outgrowth and excitatory synapse development.Methods and Results: In the present study, we investigated a single patient segregating Neurodevelopmental disorder. SLITRK2 associated significant neuropsychological issues inherited in a rare X-linked fashion have recently been reported. Whole-exome sequencing and data analysis revealed a novel nonsense variant [c.789T>A; p.(Cys263*); NM_032539.5; NP_115928.1] in exon 5 of the SLITRK2 gene (MIM# 300561). Three-dimensional protein modeling revealed substantial changes in the mutated SLITRK2 protein, which might lead to nonsense-medicated decay.Conclusion: This study confirms the role of SLITRK2 in neuronal development and highlights the importance of including the SLITRK2 gene in the screening of individuals presenting neurodevelopmental disorders.

BackgroundDevelopmental and epileptic encephalopathies (DEEs) signify a group of heterogeneous neurodevelopmental disorder associated with early-onset seizures accompanied by developmental delay, hypotonia, mild to severe intellectual disability, and developmental regression. Variants in the DNM1 gene have been associated with autosomal dominant DEE type 31A and autosomal recessive DEE type 31B.MethodsIn the current study, a consanguineous Pakistani family consisting of a proband (IV-2) was clinically evaluated and genetically analyzed manifesting in severe neurodevelopmental phenotypes. WES followed by Sanger sequencing was performed to identify the disease-causing variant. Furthermore, 3D protein modeling and dynamic simulation of wild-type and mutant proteins along with reverse transcriptase (RT)–based mRNA expression were checked using standard methods.ResultsData analysis of WES revealed a novel homozygous non-sense variant (c.1402G>T; p. Glu468*) in exon 11 of the DNM1 gene that was predicted as pathogenic class I. Variants in the DNM1 gene have been associated with DEE types 31A and B. Different bioinformatics prediction tools and American College of Medical Genetics guidelines were used to verify the identified variant. Sanger sequencing was used to validate the disease-causing variant. Our approach validated the pathogenesis of the variant as a cause of heterogeneous neurodevelopmental disorders. In addition, 3D protein modeling showed that the mutant protein would lose most of the amino acids and might not perform the proper function if the surveillance non-sense-mediated decay mechanism was skipped. Molecular dynamics analysis showed varied trajectories of wild-type and mutant DNM1 proteins in terms of root mean square deviation, root mean square fluctuation and radius of gyration. Similarly, RT-qPCR revealed a substantial reduction of the DNM1 gene in the index patient.ConclusionOur finding further confirms the association of homozygous, loss-of-function variants in DNM1 associated with DEE type 31B. The study expands the genotypic and phenotypic spectrum of pathogenic DNM1 variants related to DNM1-associated pathogenesis.

The COVID-19 pandemic has impacted the world population adversely, posing a threat to human health. In the past few years, various strains of SARS-CoV-2, each with different mutations in its structure, have impacted human health in negative ways. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) mutations influence the virulence, antibody evasion, and Angiotensin-converting enzyme 2 (ACE2) affinity of the virus. These mutations are essential to understanding how a new strain of SARS-CoV-2 has changed and its possible effects on the human body. This review provides an insight into the spike mutations of SARS-CoV-2 variants. As the current scientific data offer a scattered outlook on the various type of mutations, we aimed to categorize the mutations of Beta (B.1.351), Gamma (P.1), Delta (B.1.612.2), and Omicron (B.1.1.529) systematically according to their location in the subunit 1 (S1) and subunit 2 (S2) domains and summarized their consequences as a result. We also compared the miscellany of mutations that have emerged in all four variants to date. The comparison shows that mutations such as D614G and N501Y have emerged in all four variants of concern and that all four variants have multiple mutations within the N-terminal domain (NTD), as in the case of the Delta variant. Other mutations are scattered in the receptor binding domain (RBD) and subdomain 2 (SD2) of the S1 domain. Mutations in RBD or NTD are often associated with antibody evasion. Few mutations lie in the S2 domain in the Beta, Gamma, and Delta variants. However, in the Omicron variant many mutations occupy the S2 domain, hinting towards a much more evasive virus.

Vibrio vulnificus is a rod shape, Gram-negative bacterium that causes sepsis (with a greater than 50% mortality rate), necrotizing fasciitis, gastroenteritis, skin, and soft tissue infection, wound infection, peritonitis, meningitis, pneumonia, keratitis, and arthritis. Based on pathogenicity V. vulnificus is categorized into three biotypes. Type 1 and type 3 cause diseases in humans while biotype 2 causes diseases in eel and fish. Due to indiscriminate use of antibiotics V. vulnificus has developed resistance to many antibiotics so curing is dramatically a challenge. V. vulnificus is resistant to cefazolin, streptomycin, tetracycline, aztreonam, tobramycin, cefepime, and gentamycin. Subtractive genome analysis is the most effective method for drug target identification. The method is based on the subtraction of homologous proteins from both pathogen and host. By this process set of proteins present only in the pathogen and perform essential functions in the pathogen can be identified. The entire proteome of Vibrio vulnificus strain ATCC 27562 was reduced step by step to a single protein predicted as the drug target. AlphaFold2 is one of the applications of deep learning algorithms in biomedicine and is correctly considered the game changer in the field of structural biology. Accuracy and speed are the major strength of AlphaFold2. In the PDB database, the crystal structure of the predicted drug target was not present, therefore the Colab notebook was used to predict the 3D structure by the AlphaFold2, and subsequently, the predicted model was validated. Potent inhibitors against the new target were predicted by virtual screening and molecular docking study. The most stable compound ZINC01318774 tightly attaches to the binding pocket of bisphosphoglycerate-independent phosphoglycerate mutase. The time-dependent molecular dynamics simulation revealed compound ZINC01318774 was superior as compared to the standard drug tetracycline in terms of stability. The availability of V. vulnificus strain ATCC 27562 has allowed in silico identification of drug target which will provide a base for the discovery of specific therapeutic targets against Vibrio vulnificus.

IntroductionMonkeypox is a zoonotic disease caused by brick-shaped enveloped monkeypox (Mpox) virus that belongs to the family of ancient viruses known as Poxviridae. Subsequently, the viruses have been reported in various countries. The virus is transmitted by respiratory droplets, skin lesions, and infected body fluids. The infected patients experience fluid-filled blisters, maculopapular rash, myalgia, and fever. Due to the lack of effective drugs or vaccines, there is a need to identify the most potent and effective drugs to reduce the spread of monkeypox. The current study aimed to use computational methods to quickly identify potentially effective drugs against the Mpox virus.MethodsIn our study, the Mpox protein thymidylate kinase (A48R) was targeted because it is a unique drug target. We screened a library of 9000 FDA-approved compounds of the DrugBank database by using various in silico approaches, such as molecular docking and molecular dynamic (MD) simulation.ResultsBased on docking score and interaction analysis, compounds DB12380, DB13276, DB13276, DB11740, DB14675, DB11978, DB08526, DB06573, DB15796, DB08223, DB11736, DB16250, and DB16335 were predicted as the most potent. To examine the dynamic behavior and stability of the docked complexes, three compounds—DB16335, DB15796, and DB16250 —along with the Apo state were simulated for 300ns. The results revealed that compound DB16335 revealed the best docking score (-9.57 kcal/mol) against the Mpox protein thymidylate kinase.DiscussionAdditionally, during the 300 ns MD simulation period, thymidylate kinase DB16335 showed great stability. Further, in vitro and in vivo study is recommended for the final predicted compounds.

The new coronavirus SARS-COV-2, which emerged in late 2019 from Wuhan city of China was regarded as causing agent of the COVID-19 pandemic. The primary protease which is also known by various synonymous i.e., main protease, 3-Chymotrypsin-like protease (3CLPRO) has a vital role in the replication of the virus, which can be used as a potential drug target. The current study aimed to identify novel phytochemical therapeutics for 3CLPRO by machine learning-based virtual screening. A total of 4,000 phytochemicals were collected from deep literature surveys and various other sources. The 2D structures of these phytochemicals were retrieved from the PubChem database, and with the use of a molecular operating environment, 2D descriptors were calculated. Machine learning-based virtual screening was performed to predict the active phytochemicals against the SARS-CoV-2 3CLPRO. Random forest achieved 98% accuracy on the train and test set among the different machine learning algorithms. Random forest model was used to screen 4,000 phytochemicals which leads to the identification of 26 inhibitors against the 3CLPRO. These hits were then docked into the active site of 3CLPRO. Based on docking scores and protein-ligand interactions, MD simulations have been performed using 100 ns for the top 5 novel inhibitors, ivermectin, and the APO state of 3CLPRO. The post-dynamic analysis i.e,. Root means square deviation (RMSD), Root mean square fluctuation analysis (RMSF), and MM-GBSA analysis reveal that our newly identified phytochemicals form significant interactions in the binding pocket of 3CLPRO and form stable complexes, indicating that these phytochemicals could be used as potential antagonists for SARS-COV-2.

Recent methods for visual tracking exploit a multitude of information obtained from combinations of handcrafted and/or deep features. However, the response maps derived from these feature combinations are often fused using simple strategies such as winner-takes-all or weighted sum approaches. Although some efficient fusion methods have also been proposed, these methods still do not leverage the individual strengths of the different features being fused. In the current work, we propose a novel information fusion strategy comprising a common low-rank subspace for the fusion of different types of features and tracker responses. Firstly, we interpret the response maps as smoothly varying functions which can be efficiently represented using individual low-rank matrices, thus removing high frequency noise and sparse artifacts. Secondly, we estimate a common low-rank subspace which is constrained to remain close to each individual low-rank subspace resulting in an efficient fusion strategy. The proposed algorithm achieves good performance by integrating the information contained in heterogeneous feature types. We demonstrate the efficiency of our algorithm using several combinations of features as well as correlation filter and end-to-end deep trackers. The proposed common subspace fusion algorithm is generic and can be used to efficiently fuse the response maps of varying types of feature representations as well as trackers. Extensive experiments on several tracking benchmarks including OTB100, TC128, VOT-ST 2018, VOT-LT 2018, UAV123, GOT-10K and LaSoT have demonstrated significant performance improvements compared to many SOTA tracking methods.

Parkinson’s disease (PD) is the second most common neurodegenerative disease, with symptoms such as tremor, bradykinesia with rigidity, and depression appearing in the late stage of life. The key hallmark of PD is the loss or death of dopaminergic neurons in the region substantia nigra pars compacta. Neuroinflammation plays a key role in the etiology of PD, and the contribution of immunity-related events spurred the researchers to identify anti-inflammatory agents for the treatment of PD. Neuroinflammation-based biomarkers have been identified for diagnosing PD, and many cellular and animal models have been used to explain the underlying mechanism; however, the specific cause of neuroinflammation remains uncertain, and more research is underway. So far, microglia and astrocyte dysregulation has been reported in PD. Patients with PD develop neural toxicity, inflammation, and inclusion bodies due to activated microglia and a-synuclein–induced astrocyte conversion into A1 astrocytes. Major phenotypes of PD appear in the late stage of life, so there is a need to identify key early-stage biomarkers for proper management and diagnosis. Studies are under way to identify key neuroinflammation-based biomarkers for early detection of PD. This review uses a constructive analysis approach by studying and analyzing different research studies focused on the role of neuroinflammation in PD. The review summarizes microglia, astrocyte dysfunction, neuroinflammation, and key biomarkers in PD. An approach that incorporates multiple biomarkers could provide more reliable diagnosis of PD.

This article addresses methodologies for remote sensing of ocean Chlorophyll-a (Chl-a), with emphasis on the Barents Sea. We aim at improving the monitoring capacity by integrating in situ Chl-a observations and optical remote sensing to locally train machine learning (ML) models. For this purpose, in situ measurements of Chl-a ranging from 0.014–10.81 mg/m$^{3}$, collected for the years 2016–2018, were used to train and validate models. To accurately estimate Chl-a, we propose to use additional information on pigment content within the productive column by matching the depth-integrated Chl-a concentrations with the satellite data. Using the optical images captured by the multispectral imager instrument on Sentinel-2 and the in situ measurements, a new spatial window-based match-up dataset creation method is proposed to increase the number of match-ups and hence improve the training of the ML models. The match-ups are then filtered to eliminate erroneous samples based on the spectral distribution of the remotely sensed reflectance. In addition, we design and implement a neural network model dubbed as the ocean color net (OCN), that has performed better than existing ML-based techniques, including the Gaussian process Regression (GPR), regionally tuned empirical techniques, including the ocean color (OC3) algorithm and the spectral band ratios, as well as the globally trained Case-2 regional/coast colour (C2RCC) processing chain model C2RCC-networks. The proposed OCN model achieved reduced mean absolute error compared to the GPR by 5.2%, C2RCC by 51.7%, OC3 by 22.6%, and spectral band ratios by 29%. Moreover, the proposed spatial window and depth-integrated match-up creation techniques improved the performance of the proposed OCN by 57%, GPR by 41.9%, OC3 by 5.3%, and spectral band ratio method by 24% in terms of RMSE compared to the conventional match-up selection approach.

Optic nerve damage is a common cause of blindness. Optic nerve injury is often accompanied by fundus vascular disease, retinal ganglion cell apoptosis, and changes in retinal thickness. These changes can cause alterations in protein expression within neurons in the retina. Proteomics analysis offers conclusive evidence to decode a biological system. Furthermore, animal models of optic nerve injury made it possible to gain insight into pathological mechanisms, therapeutic targets, and effective treatment of such injuries. Proteomics takes the proteome as the research object and studies protein changes in cells and tissues. At present, a variety of proteomic analysis methods have been widely used in the research of optic nerve injury diseases. This review summarizes the application of proteomic research in optic nerve injury diseases and animal models of optic nerve injury. Additionally, differentially expressed proteins are summarized and analyzed. Various optic nerve injuries, including those associated with different etiologies, are discussed along with their potential therapeutic targets and future directions.

Background and Objective: Mutations in the CYB5R3 gene cause reduced NADH-dependent cytochrome b5 reductase enzyme function and consequently lead to recessive congenital methemoglobinemia (RCM). RCM exists as RCM type I (RCM1) and RCM type II (RCM2). RCM1 leads to higher methemoglobin levels causing only cyanosis, while in RCM2, neurological complications are also present along with cyanosis. Materials and Methods: In the current study, a consanguineous Pakistani family with three individuals showing clinical manifestations of cyanosis, chest pain radiating to the left arm, dyspnea, orthopnea, and hemoptysis was studied. Following clinical assessment, a search for the causative gene was performed using whole exome sequencing (WES) and Sanger sequencing. Various variant effect prediction tools and ACMG criteria were applied to interpret the pathogenicity of the prioritized variants. Molecular dynamic simulation studies of wild and mutant systems were performed to determine the stability of the mutant CYB5R3 protein. Results: Data analysis of WES revealed a novel homozygous missense variant NM_001171660.2: c.670A > T: NP_001165131.1: p.(Ile224Phe) in exon 8 of the CYB5R3 gene located on chromosome 22q13.2. Sanger sequencing validated the segregation of the identified variant with the disease phenotype within the family. Bioinformatics prediction tools and ACMG guidelines predicted the identified variant p.(Ile224Phe) as disease-causing and likely pathogenic, respectively. Molecular dynamics study revealed that the variant p.(Ile224Phe) in the CYB5R3 resides in the NADH domain of the protein, the aberrant function of which is detrimental. Conclusions: The present study expanded the variant spectrum of the CYB5R3 gene. This will facilitate genetic counselling of the same and other similar families carrying mutations in the CYB5R3 gene.

Infection of hepatitis C (HCV) is a major threat to human health throughout the world. The current therapy program suffers from restricted efficiency and low tolerance, and there is serious demand frr novel medication. NS3/4A protease is observed to be very effective target for the treatment of HCV. A data set of the already reported HCV NS3/4A protease inhibitors was first docked into the NS3/4A protease (PDB ID: 4A92A) active sites of both protease and helicase sites for calculating the docking score, binding affinity, binding mode, and solvation energy. Then the data set of these reported inhibitors was used in a computer-based program “RECAP Analyses” implemented in MOE to fragment every molecule in the subset according to simple retrosynthetic analysis rules. The RECAP analysis fragments were then used in another computer-based program “RECAP Synthesis” to randomly recombine and generate synthetically reasonable novel chemical structures. The novel chemical structures thus produced were then docked against HCV NS3/4A. After a thorough validation of all undertaken steps, based on Lipinski’s rule of five, docking score, binding affinity, solvation energy, and Van der Waal’s interactions with HCV NS3/4A, 12 novel chemical structures were identified as inhibitors of HCV NS3/4A. The novel structures thus designed are hoped to play a key role in the development of new effective inhibitors of HCV.

Among various physiological signal acquisition methods for the study of the human brain, EEG (Electroencephalography) is more effective. EEG provides a convenient, non-intrusive, and accurate way of capturing brain signals in multiple channels at fine temporal resolution. We propose an ensemble learning algorithm for automatically computing the most discriminative subset of EEG channels for internal emotion recognition. Our method describes an EEG channel using kernel-based representations computed from the training EEG recordings. For ensemble learning, we formulate a graph embedding linear discriminant objective function using the kernel representations. The objective function is efficiently solved via sparse non-negative principal component analysis and the final classifier is learned using the sparse projection coefficients. Our algorithm is useful in reducing the amount of data while improving computational efficiency and classification accuracy at the same time. The experiments on publicly available EEG dataset demonstrate the superiority of the proposed algorithm over the compared methods.

Among various palmprint identification methods proposed in the literature, sparse representation for classification (SRC) is very attractive offering high accuracy. Although SRC has good discriminative ability, its performance strongly depends on the quality of the training data. In particular, SRC suffers from two major problems: lack of training samples per class and large intra-class variations. In fact, palmprint images not only contain identity information but they also have other information, such as illumination and geometrical distortions due to the unconstrained conditions and the movement of the hand. In this case, the sparse representation assumption may not hold well in the original space since samples from different classes may be considered from the same class. This paper aims to enhance palmprint identification performance through SRC by proposing a simple yet efficient method based on an ensemble of sparse representations through an ensemble of discriminative dictionaries satisfying SRC assumption. The ensemble learning has the advantage to reduce the sensitivity due to the limited size of the training data and is performed based on random subspace sampling over 2D-PCA space while keeping the image inherent structure and information. In order to obtain discriminative dictionaries satisfying SRC assumption, a new space is learned by minimizing and maximizing the intra-class and inter-class variations using 2D-LDA. Extensive experiments are conducted on two publicly available palmprint data sets: multispectral and PolyU. Obtained results showed very promising results compared with both state-of-the-art holistic and coding methods. Besides these findings, we provide an empirical analysis of the parameters involved in the proposed technique to guide the neophyte.

We propose novel multi-order statistical descriptors which can be used for high speed object classification or face recognition from videos or image sets. We represent each gallery set with a global second-order statistic which captures correlated global variations in all feature directions as well as the common set structure. A lightweight descriptor is then constructed by efficiently compacting the second-order statistic using Cholesky decomposition. We then enrich the descriptor with the first-order statistic of the gallery set to further enhance the representation power. By projecting the descriptor into a low-dimensional discriminant subspace, we obtain further dimensionality reduction, while the discrimination power of the proposed representation is still preserved. Therefore, our method represents a complex image set by a single descriptor having significantly reduced dimensionality. We apply the proposed algorithm on image set and video-based face and periocular biometric identification, object category recognition, and hand gesture recognition. Experiments on six benchmark data sets validate that the proposed method achieves significantly better classification accuracy with lower computational complexity than the existing techniques. The proposed compact representations can be used for real-time object classification and face recognition in videos.

Multimedia learning is the process of building mental representation from words associated with images. Due to the intuitiveness and vividness of visual illustration, many texts to picture systems have been proposed. However, we observe some common limitations in the existing systems, such as the retrieved pictures may not be suitable for educational purposes. Also, finding pedagogic illustrations still requires manual work, which is difficult and time-consuming. The commonly used systems based on the best keyword selection and the best sentence selection may suffer from loss of information. In this paper, we present an Arabic multimedia text-to-picture mobile learning system that is based on conceptual graph matching. Using a knowledge base, a conceptual graph is built from the text accompanied with the pictures in the multimedia repository as well as for the text entered by the user. Based on the matching scores of both conceptual graphs, matched pictures are assigned relative rankings. The proposed system demonstrated its effectiveness in the domain of Arabic stories, however, it can be easily shifted to any educational domain to yield pedagogical illustrations for organizational or institutional needs. Comparisons with the current state-of-the-art systems, based on the best keyword selection and the best sentence selection techniques, have demonstrated significant improvements in the performance. In addition, to facilitate educational needs, conceptual graph visualization and visual illustrative assessment modules are also developed. The conceptual graph visualization enables learners to discover relationships between words, and the visual illustrative assessment allows the system to automatically assess the performance of a learner. The profound user studies demonstrated the efficiency of the proposed multimedia learning system.

CNN-based trackers, especially those based on Siamese networks, have recently attracted considerable attention because of their relatively good performance and low computational cost. For many Siamese trackers, learning a generic object model from a large-scale dataset is still a challenging task. In the current study, we introduce input noise as regularization in the training data to improve generalization of the learned model. We propose an Input-Regularized Channel Attentional Siamese (IRCA-Siam) tracker which exhibits improved generalization compared to the current state-of-the-art trackers. In particular, we exploit offline learning by introducing additive noise for input data augmentation to mitigate the overfitting problem. We propose feature fusion from noisy and clean input channels which improves the target localization. Channel attention integrated with our framework helps finding more useful target features resulting in further performance improvement. Our proposed IRCA-Siam enhances the discrimination of the tracker/background and improves fault tolerance and generalization. An extensive experimental evaluation on six benchmark datasets including OTB2013, OTB2015, TC128, UAV123, VOT2016 and VOT2017 demonstrate superior performance of the proposed IRCA-Siam tracker compared to the 30 existing state-of-the-art trackers.

We propose to improve the visual object tracking by introducing a soft mask based low-level feature fusion technique. The proposed technique is further strengthened by integrating channel and spatial attention mechanisms. The proposed approach is integrated within a Siamese framework to demonstrate its effectiveness for visual object tracking. The proposed soft mask is used to give more importance to the target regions as compared to the other regions to enable effective target feature representation and to increase discriminative power. The low-level feature fusion improves the tracker robustness against distractors. The channel attention is used to identify more discriminative channels for better target representation. The spatial attention complements the soft mask based approach to better localize the target objects in challenging tracking scenarios. We evaluated our proposed approach over five publicly available benchmark datasets and performed extensive comparisons with 39 state-of-the-art tracking algorithms. The proposed tracker demonstrates excellent performance compared to the existing state-of-the-art trackers.

Fall induced damages are serious incidences for aged as well as young persons. A real-time automatic and accurate fall detection system can play a vital role in timely medication care which will ultimately help to decrease the damages and complications. In this paper, we propose a fast and more accurate real-time system which can detect people falling in videos captured by surveillance cameras. Novel temporal and spatial variance-based features are proposed which comprise the discriminatory motion, geometric orientation and location of the person. These features are used along with ensemble learning strategy of boosting with J48 and Adaboost classifiers. Experiments have been conducted on publicly available standard datasets including Multiple Cameras Fall (with 2 classes and 3 classes) and UR Fall Detection achieving percentage accuracies of 99.2, 99.25 and 99.0, respectively. Comparisons with nine state-of-the-art methods demonstrate the effectiveness of the proposed approach on both datasets.