Search

Industrial developments and consumption of massive amount of fossil fuels, vehicle pollution, and other calamities upsurges the AQI (Air Quality Index) of major cities in a drastic manner. Owing to these factors, it is important to take proactive measures for reducing the air pollution in order to avoid life- threatening consequence. Therefore, prediction of air quality is significant for improving the health of living beings as highly polluted regions have a higher concentration of pollutants mixed in the air, affecting the respiratory system and reducing the lifetime. To control pollution, AQI is used as a measure for estimating the pollutant content in the air. Even though many existing techniques have predicted AQI, enhancement is required in prediction algorithms with minimized loss. To address the challenges in traditional algorithms, the proposed smart cities-based AQI prediction intends to utilize the proposed regression algorithm in the dataset, namely Air- Quality-Data, which collected harmful pollutants on an hourly and daily basis from multiple cities in India between 2015 to 2020. To achieve prediction efficiency with reduced loss, pre-processing of input data is being performed using Deep GAN (Generative Adversarial Network). It performs the imputation of data in place of missing values to improve accurate prediction. Additionally, feature scaling normalizes independent real-data features to a fixed scale. With the processed data, regression is done through modified Stacked Attention GRU with KL divergence, which predicts Ernakulam, Chennai and Ahmedabad cities with higher, medium, and low levels of AQI in India. The performance of the proposed regression algorithm is measured using metrics such as MAE (Mean Absolute Error), MSE (Mean Square Error), R2 (Coefficient of determination), MAPE (Mean Absolute Percentage Error), and RMSE (Root Mean Square Error) and better MAE, MSE, R2, MAPE and RMSE obtained by the model is 0.1013, 0.0134, 0.9479, 0.1152 and 0.1156. Internal assessment and comparative analysis performed with existing regression algorithms exhibit lower loss values obtained from the present research, which determines the efficacy of the proposed model. [ABSTRACT FROM AUTHOR]

Tamil is a language that has the most extended history and is a conventional language of India. It has antique origins and a distinct tradition. A study reveals that at the beginning of the twenty-first century, more than 66 million people spoke Tamil. In the present time, image synthesis from text emerged as a promising advancement in computer vision applications. The research work done so far in intelligent systems is trained in universal language but still has not achieved the desired development level in regional languages. Regional languages have a greater scope for developing applications and will enhance more research areas to be explored, ruling out the barrier. The current work using Auto Encoders failed at the point of providing vivid information along with essential descriptions of the synthesised images. The work aims to generate embedding vectors using a language model headed by image synthesis using GAN (Generative Adversarial Network) architecture. The proposed method is divided into two stages: designing a language model TBERTBASECASE model for generating embedding vectors. Synthesising images using Generative Adversarial Network called BASEGAN, the resolution has been improved through two-stage architecture named HYBRID SUPER RESOLUTION GAN. The work uses Oxford-102 and CUB-200 datasets. The framework efficiency has been measured using F1 Score, Fréchet inception distance (FID), and Inception Score (IS). Language and image synthesis architecture proposed can bridge the gap between the research ideas in regional languages. [ABSTRACT FROM AUTHOR]

Automatic road crack detection plays a major role in developing an intelligent transportation system. The traditional approach of in-situ inspection is expensive and requires more man-power. In-order to solve this problem, a novel approach for automatic road crack segmentation was developed using Stack Generative adversarial network Discriminator-U-Network (SGD-U-Network). We have collected 19 300 crack and non-crack images (MIT-CHN-ORR dataset) from the Outer Ring Road of Chennai, TamilNadu, India. The MIT-CHN-ORR dataset was initially pre-processed using traditional image processing techniques for ground truth image generation. A stage-I and stage-II stack Generative Adversarial Network (GAN) model was introduced for generating high-resolution non-crack images. Then, the extracted features from Stack GAN Discriminator of stage II (SGD2) was concatenated with every level of expansion path in SGD-U-Network for segmenting the crack regions of the input crack images. Also, multi-feature-based classifier was developed using the features extracted from SGD2 and the bottleneck layer of SGD-U-Network. Our proposed model was implemented on MIT-CHN-ORR dataset and also analyzed our model performance using other existing benchmark datasets. The experimental analysis showcased that the proposed method outperformed the other state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

Many people and children in the world go missing every day, which the police and local authorities must deal with. The police and municipal authorities deal with many missing persons and kid cases every day. In 2020, National Crime Records Bureau (NCRB) listed around 300 thousand missing persons in India. The general process normally deals with investigation, which requires experience and can be time-consuming if the missing person is migrated. The overall procedure usually entails inquiry, which is time-consuming and requires experience. Our research intends to create an AI system that can identify and evaluate a person's facial features and then compare them to a face in the database, even if they are located years after going missing. Using the Yale Face Dataset on Kaggle containing over 70,000 images for the model, the training of such large dataset improved the model for the Eigenface as well as the GAN algorithm with the use of OpenCV and HaarCascade library helped improve and enhance the accuracy with the result of near 99. [ABSTRACT FROM AUTHOR]

• Generative Adversarial Network for inverting Moho architecture from observed gravity anomalies. • Spherical prism based forward gravity modelling for high accuracy. • Near realistic Moho topography generation using FFT filtering technique for training the network. • Fast and accurate estimation of high-resolution Moho without prior knowledge of inversion algorithms. Accurate estimation of Moho topography plays a crucial role in understanding Earth's structure, geodynamic processes, and resource exploration. This study presents a novel approach that utilizes conditional Generative Adversarial Networks (cGAN) to reveal Moho topography based on observed gravity anomalies. Synthetic training datasets of Moho topography were generated using the FFT filtering method due to the scarcity of true datasets. Spherical prism-based forward gravity modeling was employed to evaluate the resulting gravity anomalies. We compared the performance of our developed deep learning algorithm cGAN (conditional Generative Adversarial Networks) with a traditional inversion technique using various synthetic datasets, and a real case study in southern peninsular India, a geologically diverse region comprising ancient continental tectonic blocks. Bott's inversion scheme was employed as a verification method for the Moho surface estimation using the presented deep learning model. Using spherical prism-based forward gravity modeling, observed gravity anomalies were corrected for multiple factors such as topography, bathymetry, sediments, crustal heterogeneities, and mantle heterogeneities. By removing these effects, we isolated the gravity contribution solely related to pure Moho undulation. The mean Moho depth and density contrast between the crust and mantle were derived from seismic constraints for improving estimation accuracy. The findings demonstrate the potential of the cGAN and spherical prism-based gravity modeling approach in accurately estimating the Moho topography, offering insights into Earth's subsurface structures and enhancing our understanding of geodynamic processes and resource exploration efforts. [ABSTRACT FROM AUTHOR]

Early rainfall prediction is very important to ensure the economic balance of any agriculture-dominated country, such as India. Deep learning has recently received considerable attention for predicting rainfall using historical data. However, rainfall depends on multiple factors like ElNino, rising sea levels, warming of the Atlantic Ocean, etc. The changing climate has also significantly increased rainfall variability. The current deep-learning approaches are either based on a limited set of predictors or do not utilize the time-series properties for the recurrence of similar climate conditions every 12 months. Besides, they perform poorly in predicting accurate rainfall estimates for the diverse areas of India due to limited data availability across various regions. Therefore, this paper presents a groundbreaking solution to address the challenge of limited data by introducing RfGan (Rainfall GAN), a novel variant of generative adversarial networks. RfGan aims to fulfill the crucial need for abundant data to ensure the effectiveness of model training. Moreover, a novel rainfall prediction method, RfGanNet, has been introduced that leverages the strengths of the proposed RfGan for data augmentation, long-short-term memory, and convolutional neural networks for rainfall prediction. The performance of RfGanNet has been validated for India and its clustered regions in terms of nine parameters: precision, recall, specificity, Matthews correlation coefficient, accuracy, F1-score, mean absolute error, mean square error and computation time. The experimental results of RfGan are compared against original and synthetic datasets generated from state-of-the-art methods, namely tabular gan (TGAN), Wasserstein gan (WGAN), and conditional tabular gan (CTGAN). The experimental results indicate that the proposed RfNet model, when trained using RfGan-generated data, surpassed all the other considered methods by achieving 97.22 % accuracy. Moreover, RfNet has been compared with other related deep Learning approaches to validate its efficacy. Additionally, an ablation study investigating various LSTM–CNN combinations has been conducted. The findings of the study indicate that oceanic and climate-related predictors have a noteworthy influence on Indian rainfall. Moreover, the study highlights the suitability of combining Bidirectional LSTM, CNN, and multiple predictors for accurate rainfall prediction. • Novel variant of generative adversarial network- RfGan is proposed. • Different combinations of LSTM and CNN for rainfall prediction are tested. • Novel classifier — RfNet is tested on different augmented datasets. • K-means clustered regions of India have been identified. [ABSTRACT FROM AUTHOR]

Noise pollution, a growing problem in urban cities, causes severe and long-term health problems. Vehicles on the road and ceaseless honking add to incessant vehicular assault and are the main reasons behind increasing noise pollution. However, analyzing and characterization of honking is not an easy task, and it varies with road traffic, vehicle types, speed, etc. In this paper, we propose a framework De H onk that detects vehicular honks from traffic noise and analyzes various honking features to build a honking profile using only audio samples, which can be further used to infer the context of a particular location, design micro-services, etc. However, the existing methods used only fast fourier transform(FFT) or mel-frequency cepstral coefficients(MFCC) in honk identification, the accuracy lies between 40% to 60% in the presence of ambient noise. Similarly, we failed to obtain good accuracy in recognizing honks at noisy environments by applying existing sound classification models. In the proposed work, we have modeled the sound signal as a spectrogram image to detect vehicular honks. We have collected audio data by traveling 461km of road from various spatio-temporal locations in Durgapur, a sub-urban city in India. Furthermore, the generated spectrogram images are pre-processed, labeled, and fed into various deep learning models for training and testing. In this study, convolutional neural network(CNN), visual geometry group with a very deep convolutional network(VGGNets), efficient convolutional neural networks for mobile vision applications(MobileNet), residual neural network(ResNets), and Inception V3 models are used to train the system to identify honks and honk related features. Experiment results demonstrate that the ResNet152 model has achieved higher accuracy (97.69%) in honk detection as compared to other models. Furthermore, as a micro-service, a honk-aware route recommendation system is developed based on the derived honking features. [Display omitted] • Data labeling is done using semi-supervised generative adversarial networks (SGAN). • Different deep learning is used to identify vehicular honks and its feature. • Honk-aware route recommendation system based on our derive context. [ABSTRACT FROM AUTHOR]