Search

Monitoring the healthy development of a fetus requires accurate and timely identification of different maternal–fetal structures as they grow. To facilitate this objective in an automated fashion, we propose a deep-learning-based image classification architecture called the COMFormer to classify maternal–fetal and brain anatomical structures present in 2-D fetal ultrasound (US) images. The proposed architecture classifies the two subcategories separately: maternal–fetal (abdomen, brain, femur, thorax, mother’s cervix (MC), and others) and brain anatomical structures [trans-thalamic (TT), trans-cerebellum (TC), trans-ventricular (TV), and non-brain (NB)]. Our proposed architecture relies on a transformer-based approach that leverages spatial and global features using a newly designed residual cross-variance attention block. This block introduces an advanced cross-covariance attention (XCA) mechanism to capture a long-range representation from the input using spatial (e.g., shape, texture, intensity) and global features. To build COMFormer, we used a large publicly available dataset (BCNatal) consisting of 12 400 images from 1792 subjects. Experimental results prove that COMFormer outperforms the recent CNN and transformer-based models by achieving 95.64% and 96.33% classification accuracy on maternal–fetal and brain anatomy, respectively.

The access to knowledge is an important requirement for advancement of scientific and technological research and development of a country. Availability of resources is a crucial bottleneck for universities and colleges in developing countries. This leads to frequent use of pirate access sites like Sci-Hub by researchers. For instance, India has more than 900 universities and 40,000 colleges, and Africa has more than 1200 universities. Only a few of these institutions would have access to most of the research journals their scholars require. It is in this context that we tried to find out if there exist some resources which can provide links to open and free to download versions of scientific papers. Google Scholar, a heavily used resource for research article searches, is explored to see how effective it is in providing links to open access freely downloadable copies of scientific articles. The complete set of global scientific publications for the year 2016 are computationally analyzed through a web-mining approach, as an example, to see if Google Scholar is able to point to freely downloadable open text versions of scientific articles. Results show that Google Scholar points to full-text sources for about 69% of the articles queried, with about 43% of the articles having openly accessible full-texts. The results, thus, indicate that Google Scholar can be a useful tool for locating open access full-text versions of close to about half of the scientific articles of the world, which has special significance for under-developed and developing countries.

Significant efforts have been made to monitor human activity, although it remains a challenging area for computer vision research. This paper has introduced a framework to identify the most common types of video surveillance activities. The proposed framework consists of three consecutive modules: (i) human detection by background subtraction, (ii) retrieval of uniform and rotation invariant local binary pattern (LBP) feature, and (iii) identification of human activities with a support vector machine (SVM) multiclass classifier. This framework provides a consistent view of the human actions that look at multiple subjects from different views. In addition to this, uniform patterns provide better performance in discriminating human activities. A multiclass SVM is used for classification of human activities. SVM classifier is set and trained to achieve the better efficiency by selecting the appropriate feature before it is integrated. Weizmann's Multiview dataset, CASIA dataset and IXMAS dataset confirm the high efficiency and better robustness of the proposed framework.

Agriculture is a vital sector in developing nations such as India, and the use of autonomous vehicles and Internet of Things (IoT) technology has the potential to revolutionize farming practices. Unmanned Aerial Vehicles (UAVs) are becoming increasingly important in agriculture, as they can provide valuable data for crop monitoring and pest control. In this study, we investigate the reliability of a Multi-Agent Reinforcement Learning (MARL) method for UAV field coverage. The algorithm enables a group of UAVs equipped with ground-facing cameras to learn how to provide complete coverage of an unknown Field of Interest (FoI) while minimizing camera view overlap. We test the algorithm in scenarios where the FoI and camera Field of View (FoV) are dynamically updated in the environment, to evaluate its performance under more dynamic conditions. Our results demonstrate the effectiveness and resilience of the proposed method in varying environmental conditions, highlighting its potential for Precision Agriculture (PA) applications.

In recent years, the segmentation of anatomical or pathological structures using deep learning has experienced a widespread interest in medical image analysis. Remarkably successful performance has been reported in many imaging modalities and for a variety of clinical contexts to support clinicians in computer-assisted diagnosis, therapy, or surgical planning purposes. However, despite the increasing amount of medical image segmentation challenges, there remains little consensus on which methodology performs best. Therefore, we examine in this article the numerous developments and breakthroughs brought since the rise of U-Net-inspired architectures. Especially, we focus on the technical challenges and emerging trends that the community is now focusing on, including conditional generative adversarial and cascaded networks, medical Transformers, contrastive learning, knowledge distillation, active learning, prior knowledge embedding, cross-modality learning, multistructure analysis, federated learning, or semi-supervised and self-supervised paradigms. We also suggest possible avenues to be further investigated in future research efforts.

AbstractRapid urbanization and land use transformation coupled with climate change impact are witnessed for development urban heat island worldwide and make it major environmental concerns for human health. National Capital Region (NCR) is one of the fast sprawling National Capital city of India. The aim of the present study is to explore the spatial changes in major biophysical parameters of city over last 20 years from 2000 to 2020 using satellite data. The spatial variation in the land surface temperature and its relation with Land use, NDVI and Built-up area were evaluated and discussed in the present work. The major changes were observed in minimum and maximum temperature of the city, in 2000 minimum temperature was 23.20 °C and maximum 34.85 °C, which increased to 26.31 and 39.92 °C in 2010; and 31.70 to 44.74 °C in 2020 . Normalized differential vegetation index (NDVI) decrease from 82 to 62%, Non-Vegetated area has reduced to 62% from 82% due to fast urbanization and land transformation in last two decades. Barren and fallow land also reduces about 27% in last two decades and built-up area also shown increasing trend . The spatial changes in biophysical parameter of the city clearly indicated that NDVI, NDBI and Land use/land cover charges are the main driving forces for development of urban heat island in the city and many health issues related to heat waves are already reported in the city along with many other negative impacts on natural landscape and resources. The observations from present work will be very useful scientific information to understand the causes of elevated land surface temperature and mitigation measures. Results could be used by policy makers to develop the suitable model for long term sustainability of the cities.

In this paper, a method for maximum fault coverage with minimum power dissipation, during the testing of analog and digital circuits of mixed-signal System-on-Chip (SOC) simultaneously using Genetic Algorithm, (GA) is proposed. Mixed-signal SOC consists mainly of an analog block, a digital block and a DAC/ADC (Digital to Analog Converter/Analog to Digital Converter). Due to the presence of analog and digital circuits in mixed-signal SOCs, the testing procedure is difficult from that of only analog or digital circuit testing. Here stuck at 0/1 faults are considered for digital circuits, and stuck open/short faults are considered for analog circuits. In analog testing fault modeling, fault injection and fault simulation are done. The outputs of the analog block and some independent digital signals are given to the digital block. The GA-based approach is used for power-aware ordering of test patterns considering pattern dependency on previous patterns at the input of digital block as input patterns. The effect of noise on analog test signals has also been investigated here with the area analysis of the circuit. An average of 92.43% fault coverage and 13.56% maximum power saving is achieved when this methodology is applied to ITC 97 (Analog block) and ISCAS 85 (Digital block) benchmark circuits, respectively. A trade-off between fault coverage and power dissipation has been presented.

Social media has become a platform for expressing opinions and emotions, but some people also use it to spread hate, targeting individuals, groups, communities, or countries. Therefore, there is a need to identify such content and take corrective action. During the last few years, several techniques have been developed to automatically detect and identify hate speech, offensive and abusive content from social media platforms. However, majority of the studies focused on hate speech detection in English language texts only. The non-availability of suitable datasets is a major reason for lack of research work in other languages. Hindi is one such widely spoken language where such datasets are not available. This work attempts to bridge this research gap by presenting a curated and annotated dataset for target-based hate speech (TABHATE) in the Hindi language. The suitability of the dataset is explored by applying some standard deep learning and transformer-based models for the task of hate speech detection. The experimental results obtained show that the dataset can be used for experimental work on hate speech detection of Hindi language texts.

Due to population growth and rapid urbanization in Indian cities, transportation has evolved as a critical concern affecting a large number of commuters everyday. Hence it is important for the urban planners, policymakers, and transportation authorities of India to know about the different public grievances/concerns regarding transportation. This study aims to uncover valuable information about specific transport-related complaints/grievances in Indian cities from the vast pool of user-generated content on social media platforms such as Twitter. As an initial step, we have explored the broad sentiment of commuters in six Indian metropolitan cities about the existing transportation systems, and created a dataset that broadly classify tweets into negative and positive sentiments. Next, we have identified a set of fine-grained complaints/grievances in these tweets, and thus created the first dataset containing transport-related tweets labelled into various specific complaints/grievances in a multi-label setting. To our knowledge, there is no existing dataset that labels tweets according to specific concerns raised in the posts. We apply several classification models on the dataset, for classifying transportation-related tweets into the specific complaints/grievances. We further conducted a city-wise analysis to better comprehend the specific transport-related complaints prevalent in each Indian city.

ABSTRACTThe wider adoption of AI and machine learning (ML) applications has been limited by the high costs of infrastructure and scarcity of ML experts and data scientists. To address some of these concerns, automated ML (AutoML) systems have been developed alongside cloud computing platforms to mitigate some of the constraints in the wider adoption of ML technologies, including by small and medium size organizations. In this paper, we introduce AutoML, identify some of the fundamental steps in model development, and currently available operationalizations of these systems, before concluding with an outline of potential research opportunities for IS researchers in the field.

Web applications are extensively used all over the world-wide net. These are the applications that have made our regular day to day existence easy and straightforward. We share incredibly confidential information over the web and keeping the information confidential is the most important priority of these web-based applications. These web-based applications are vulnerable to different attacks, XSS and SQL injection being one of the most perceived attacks. Attackers inject malicious code into the web application and can cause cataclysmic harm to the person without his/her consent. Cross-site scripting (XSS) and SQL injection strikes are the most generally perceived sort of attacks that a web application experience. The XSS attack is caused by executing malicious code to a web browser which runs in the server about which the user is not aware of and is capable of catastrophic disaster. Recognising a script that is malicious or benign can keep an unwanted attack from happening and can help the web application from keeping the data of the users confidential. The work accentuates on a portion of the conceptual approaches to identify the XSS script on the customer side dependent on a portion of the highlights extracted from the vindictive contents.

This paper presents an efficient and dynamic approach to a dual gating strategy in finite state machines (FSMs) to reduce the overall power dissipation. The FSMs are implemented in configurable logic blocks (CLBs) of an FPGA, and the gating criteria are derived from the self-loop information within the FSM. Fine-grained power gating may cause practical mistakes in CLB functioning, resulting in data loss. As a result, in addition to power gating, a clock gating approach is proposed to maintain data at the output of power gated CLBs. The simulations were conducted for standard MCNC benchmark circuits implemented in CLBs, designed and simulated in the 45 nm technology CADENCE tool, resulting in an overall 25.82% power savings at the cost of excess look up table (LUT) utilisation within CLBs. As the complexity of the FSM increases, the excess LUT utilisation associated with this methodology decreases.

A ruthenium(ii) complex of a chelated heteroditopic N-heterocyclic carbene ligand in combination with a phosphine ligand was uncovered to be a highly effective catalyst for the N-alkylation of diverse amides using readily available primary alcohols. A wide range of secondary amides was thus obtained in excellent yields (up to 98%) employing a low catalyst (2c) loading of 0.2 mol% and a substoichiometric amount of base. The 1H NMR and ESI-MS analyses support the participation of a N-heterocyclic carbene and phosphine supported Ru–H species in the catalytic cycle and the mechanistic studies including the deuterium labelling experiment suggest the involvement of a borrowing hydrogen protocol. Additionally, the present catalytic system was also revealed to be efficient for the selective mono-alkylation and unsymmetrical di-alkylation of 4-aminobenzamides which have not been studied before to the extent of our knowledge.

Low-cost potable water is a challenge nowadays and research are going on based on the solar thermal technique in order to bring down the cost of yield worldwide. The evacuated tubular collector is generally used to achieve higher collector efficiency. A higher rate of vaporization is achieved by integrating a number of series-connected evacuated tubes with solar still. In this work, the effect of heat transfer rate on the performance of evacuated tubular collector systems and natural evaporation and condensation-based freshwater yield in free and forced modes is reviewed. The performance of the evacuated tube is affected by varying tilt angles and an increase in the intensity of solar radiation. The overall performance of the system depends on the heat transfer rate from the absorber tube to the working fluid. The heat transfer rate can further be improved by adding various sizes of nanoparticles to working fluid and act as a heat-absorbing device during less sunshine or cloudy days. Previous finding such as operating temperature and flow behavior inside evacuated tube collector, daily freshwater yield variation with respect to depth of water inside solar still, system efficiency of the integrated system and variation in energy and exergy efficiency are presented step by step throughout the paper.

The recently launched large language models have the capability to generate text and engage in human-like conversations and question-answering. Owing to their capabilities, these models are now being widely used for a variety of purposes, ranging from question answering to writing scholarly articles. These models are producing such good outputs that it is becoming very difficult to identify what texts are written by human beings and what by these programs. This has also led to different kinds of problems such as out-of-context literature, lack of novelty in articles, and issues of plagiarism and lack of proper attribution and citations to the original texts. Therefore, there is a need for suitable computational resources for developing algorithmic approaches that can identify and discriminate between human and machine generated texts. This work contributes towards this research problem by providing a large sized curated and annotated corpus comprising of 44,162 text articles sourced from Wikipedia and ChatGPT. Some baseline models are also applied on the developed dataset and the results obtained are analyzed and discussed. The curated corpus offers a valuable resource that can be used to advance the research in this important area and thereby contribute to the responsible and ethical integration of AI language models into various fields.

Trademarks law is considered as the most pervasive amongst all the intellectual property laws because all the judgments require contemplating the imaginations of the consumers. Once the trademark has been submitted for registration, the examiners at the trademark office make sure that it is not similar to any of the previous registered trademarks. This motivated the need of an automated trademark retrieval system. This paper makes a contribution in the field of trademark image retrieval by proposing a retrieval technique that allows a flexible combination of colour, texture and shape features. Moreover the proposed technique utilises HSV colour histogram, for colour, multi resolution Gabor wavelet for texture and an integration of Zernike moments for global shape and scale invariant feature transform (SIFT) for local shape feature extraction. The results have been tested on MPEG7, MPEG trademark, WANG and self-compiled datasets. The improvement achieved in precision is 14% on MPEG7, 30% on the WANG and 38% on self-compiled dataset. Similarly, 26% improvement in average recall is achieved and around 16% when our proposed shape feature is compared with other state of the art techniques like Fourier descriptors, Hu moments, wavelet descriptors, Zernike moments and edge gradient co-occurrence matrix.

The microstructures of additively manufactured Inconel 718 consist of the primary phase (dendritic) and the secondary phases (Laves and carbides). Among the secondary phases, the Laves phase, in particular, adversely affects the mechanical performance of the material. A post-processing heat treatment is necessary to eliminate undesirable secondary phases. The objective of the present study is to eliminate long-chain Laves using heat treatment below the δ – solvus temperature and thereby establish the real-time dissolution mechanism of Laves phase. The specimens are produced using the laser powder bed fusion (LPBF) process, followed by two different heat treatment procedures based on the temperatures determined using Scheil calculation, and those below the δ – solvus temperature. While the Laves phase is invariably found in the as-printed stage, this study demonstrates the successful elimination of a long singular chain of Laves phase when heat treated below the δ – solvus temperature. Additionally, the real-time transformation and growth mechanisms of δ – phase during annealing, are effectively captured through in situheat treatment in a Transmission Electron Microscope (TEM). A re-distribution of the solute from the Laves phase to the γ – matrix is observed, with the δ – phase in the form of a needle on either side of the Laves at 900 ˚C. The in situobservations have shed important insights on the transformation mechanism of Laves phase to δ – phase in additively manufactured Inconel 718 samples.

In this paper, a new data acquisition scheme is proposed based on compressed sensing (CS) with on-demand explosion of random walks (RWs) for improving the lifespan of wireless sensor networks (WSNs) without compromising its detection capability. From the perspective of sensor networks, the lifetime largely relies upon communication efficiency and load distribution on the nodes. CS with RW is proven to be an efficient mechanism for data gathering for sparse recovery, but requires minimum $O(s \times \log(n/s))$ measurements. To lift this assumption, we have exploited the rate of change of data, which works in a similar fashion as sparsity does. Moreover, it also yields a way for opportunistic initiation of RWs which ensures explicit event detection in the target region. Therefore, an on-demand explosion-based compressed sensing (ODECS) is proposed to reduce the required measurements dynamically and limits the required measurements to a constant in case of no events. Simulation and experimental analysis show that ODECS outperforms the existing schemes in terms of data transmission, energy consumption, and lifetime by reducing communication which depends upon the probability of occurrence of event in each data acquisition round.

This paper proposes an efficient technique to develop an automated offline signature verification system that could help in crime prevention and biometric authentication systems. The technique proposed makes use of direction-based methods to compute a set of features that are taken together as a combination. The features include the geometric details of the different strokes that compose a signature and contours of the signature. It includes the two-step/three-step features, radical points, directions and transitions of strokes and contours, energy density and angles of strokes to name a few. A grid-based approach is applied to extract some of the features. Classification is done by using Support Vector Machine (SVM). Experiments are performed on a standard CEDAR database and a self-prepared database. The results speak for the efficiency of the proposed system that achieves accuracy level much better than many of the published works.

Adenosine triphosphate (ATP) is an important transmitter that mediates various biological effects via purinergic receptors (P2 receptors) in cancer. We investigated the antitumor activity of ATP-decorated and doxorubicin (DOX)-loaded mesoporous silica with biomineralization of calcium carbonate against a highly aggressive and metastatic murine lymphoma called Dalton’s lymphoma (DL). Our results suggest that this nanocomposite has unique effects with respect to the morphology and properties of calcium carbonate on the surface of the nanoparticle. DOX in the nanoparticles was prevented from quick release via the interactions of the phosphate group present on ATP and calcium carbonate. This construct is significantly tumoricidal against parental and DOX-resistant DL cells and is thus a promising candidate for applications in drug delivery. The composite nanomaterial has excellent biocompatibility with higher uptake and acts via the participation of the purinergic receptor P2X7. The nanocomposite induces significantly higher apoptosis in tumor cells compared with DOX alone. Treatment of DL-bearing mice with the construct significantly reduces tumor burden, in addition to augmenting the lifespan of tumor-bearing mice as demonstrated by a sustained healthy life of the animals and improved histopathological parameters.

Chronic myeloid leukemia (CML) is caused by chromosomal rearrangement resulting in the expression of Bcr-Abl fusion protein with deregulated Abl tyrosine kinase activity. Approved drugs – imatinib, dasatinib, nilotinib, and ponatinib – target the ATP-binding site of Abl kinase. Even though these drugs are initially effective, long-term usefulness is limited by the development of resistance. To overcome this problem, targeting the allosteric site of Abl kinase, which is remote from the ATP-binding site is found to be a useful strategy. In this study, structure-based and ligand-based virtual screening methods were applied to narrow down possible drugs (from DrugBank database) that could target the allosteric site of Abl kinase. Detailed investigations of the selected drugs in the allosteric site of Abl kinase, using molecular dynamics and steered molecular dynamics simulation shows that gefitinib, an EGFR inhibitor approved for the treatment of lung cancer, could bind effectively to the allosteric site of Bcr-Abl. More interestingly, gefitinib was found to enhance the ability of imatinib to bind at the ATP-binding site of Bcr-Abl kinase. Based on the in silicofindings, gefitinib was tested in combination with imatinib in K562 CML cell line using MTT cell proliferation assay and found to have a synergistic antiproliferative activity. Further detailed mechanistic study could help to unravel the full potential of imatinib – gefitinib combination for the treatment of CML.

In this study we present for the first time tetrahedral amorphous carbon (ta-C)a partially reduced graphene oxide (PRGO) hybrid electrode nanomaterial platform for electrochemical sensing of dopamine (DA). Graphene oxide was synthesized with the modified Hummer’s method. Before modification of ta-C by drop casting, partial reduction of the GO was carried out to improve electrochemical properties and adhesion to the ta-C thin film. A facile nitric acid treatment that slightly reoxidized the surface and modified the surface chemistry was subsequently performed to further improve the electrochemical properties of the electrodes. The largest relative increase was seen in carboxyl groups. The HNO3treatment increased the sensitivity toward DA and AA and resulted in a cathodic shift in the oxidation of AA. The fabricated hybrid electrodes were characterized with scanning electron microscopy (SEM), Raman spectroscopy, Fourier transform infrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS), X-ray absorption spectroscopy (XAS), and electrochemical impedance spectroscopy (EIS). Compared to the plain ta-C electrode the hybrid electrode was shown to exhibit superior sensitivity and selectivity toward DA in the presence of ascorbic acid (AA), enabling simultaneous sensing of AA and DA close to the physiological concentrations by cyclic voltammetry (CV) and by differential pulse voltammetry (DPV). Two linear ranges of 0–1 μM and 1–100 μM and a detection limit (S/N = 3.3) of 2.6 nM for DA were determined by means of cyclic voltammetry. Hence, the current work provides a fully CMOS-compatible carbon based hybrid nanomaterial that shows potential for in vivomeasurements of DA.

Chromium oxide is a promising material for applications where excellent corrosion resistance, high hardness, and high wear resistance are needed. However, its use is limited because of low fracture toughness. Improvement of fracture toughness of chromium oxide while maintaining its afore mentioned key properties is therefore of high interest. In this communication we study the possibility of increasing the toughness of pulsed electric current sintered (PECS) chromium oxide by the addition of graphene oxide (GO). The indentation fracture toughness was improved markedly with the addition of graphene oxide. Materials prepared by direct chemical homogenization had better fracture toughness. In composites with 10 vol.% GO piling of thin graphene oxide layers resulted in the formation of graphite layers between Cr2O3 and in carbide formation, which were observed to be the main reasons for the degradation of the mechanical properties. The distribution of graphene oxide was more homogeneous, when the GO amount was 0.1 vol.% and the formation of graphitic layers were avoided due to lesser amount of GO as well as ultrasonic treatment following the ball milling.

ABSTRACTThis paper presents our experimental work towards detecting sentiment polarity of free-form texts: first by using an ensemble of sentiment lexicons and then through a lexicon pooled machine learning classifier. In the ensemble design, we combined four different sentiment lexicons in different ways to determine sentiment polarities of different text data. The ensemble approach, however, did not achieve superior performance as initially thought. Therefore, in the second design, we tried to pool the sentiment lexicon knowledge into the machine learning classification process itself of a multinomial naive Bayes classifier. The experimental designs are evaluated on three document and two sentence datasets. The lexicon pooled approach obtains superior accuracy levels as compared to standard naive Bayes classifier as well as lexicon-based methods. Furthermore, as the amount of training data decreases, the accuracy levels of lexicon pooled machine learning classifier decays slowly as compared to standalone naive Bayes classifier. The framework presented proves useful and robust and can be extended to any classification task.

Artificial Neural Network (ANN) with adaptability and non-linearity is well suited to perform forecasting tasks. The paper proposes the systematic approach of monthly energy forecasting using Windowed Momentum Algorithm in ANN. It describes the mathematical foundations and implementation to calculate the monthly energy demand using previous three years consecutive monthly energy data and weather information. The results are compared with the standard momentum, generally used in Back Propagation Algorithm. The proposed algorithm shows good accuracy with maximum mean absolute percentage error of 1.177%.