International Journal of Next-Generation Computing

Dynamic Hand Gesture Recognition for Indian Sign Language using Integrated CNN-LSTM Architecture

Pradip Patel — 2023-11-28

Human Centered Computing is an emerging research field that aims to understand human behavior. Dynamic hand gesture recognition is one of the most recent, challenging and appealing application in this field. We have proposed one vision based system to recognize dynamic hand gestures for Indian Sign Language (ISL) in this paper. The system is built by using a unified architecture formed by combining Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM). In order to hit the shortage of a huge labeled hand gesture dataset, we have created two different CNN by retraining a well known image classification networks GoogLeNet and VGG16 using transfer learning. Frames of gesture videos are transformed into features vectors using these CNNs. As these videos are prearranged series of image frames, LSTM model have been used to join with the fully-connected layer of CNN. We have evaluated the system on three different datasets consisting of color videos with 11, 64 and 8 classes. During experiments it is found that the proposed CNN-LSTM architecture using GoogLeNet is fast and efficient having capability to achieve very high recognition rates of 93.18%, 97.50%, and 96.65% on the three datasets respectively.

Online Shoppers' Purchase Intention using Ensemble Learning Approach

Jhimli Adhikari — 2023-11-28

The customer’s purchase intention can be predicted by analyzing the history of the customers. In this study, we have analyzed the data of online shoppers for building a better prediction model to predict their purchase intention. Initially exploratory data analysis is performed on the dataset. We have used different ensemble algorithms such as Random Forest, Gradient Boosting, XGBoost and LightGBM to predict whether a customer, visiting the website of an online shop, will end up with a purchase or not. Later we have performed ensemble methods to boost up the performance of the algorithms using SMOTE to overcome class imbalance. Lastly model performance evaluation is done using parameter tuning. Study has shown that XGBoost model predicts with 89.97 % accuracy on imbalanced data, whereas Random Forest displays 93% accuracy after using SMOTE to predict the customer’s purchase intention. Moreover, XGBoost shows the highest accuracy, which is 93.54% after parameter tuning. Execution time of the models is also observed in the study.

Towards Conceptualization Of A Prototype For Quantum Database: A Complete Ecosystem

Sayantan Chakraborty — 2023-11-28

This study proposes a conceptualization of a prototype And a possibility to converge classical database and fully quantum database. This study mostly identifies the gap between this classical and quantum database and proposes a prototype that can be implemented in future products. It is a way that can be used in future industrial product development on hybrid quantum computers. The existing concept used to consider oracle as a black box in this study opens up the possibility for the quantum industry to develop the QASAM module so that we can create a fully quantum database instead of using a classical database as BlackBox.As the Toffoli gate is basically an effective NAND gate it is possible to run any algorithm theoretically in quantum computers. So we will propose a logical design for memory management for the quantum database, security enhancement model, Quantum Recovery Manager & automatic storage management model, and more for the quantum database which will ensure the quantum advantages. In this study, we will also explain the Quantum Vector Database as well as the possibility of improvement in duality quantum computing. It opens up a new scope, possibilities, and research areas in a new approach for quantum databases and duality quantum computing.

Interpolation Based Reversible Data Hiding using Pixel Intensity Classes

Abhinandan Tripathi — 2023-11-28

In this article, we suggest a new interpolation technique as well as a novel Reversible Data Hiding (RDH) approach for up scaling the actual image and concealing sensitive information within the up scaled/interpolated image. This data hiding strategy takes into account the features of the Human Visual System (HVS) when concealing the secret data in order to prevent detection of the private data even after extensive embedding. The private data bits are adaptively embedded into the picture cell based on its values in the suggested hiding strategy after grouping different pixel intensity ranges. As a result, the proposed approach can preserve the stego-visual image’s quality. According to experimental findings, the proposed interpolation approach achieves PSNRs of over 40 dB for all experimental images. The outcomes further demonstrate that the suggested data concealing strategy outperforms every other interpolation-based data hiding scheme existing in use.

Forecasting Time Series AQI Using Machine learning of Haryana Cities Using Machine Learning

Reema Gupta — 2023-11-28

In India and throughout the world, air pollution is becoming a severe worry day by day. Governments and the general public have grown more concerned about how air pollution affects human health. Consequently, it is crucial to forecast the air quality with accuracy. In this paper, Machine learning methods SVR and RFR were used to build the hybrid forecast model to predict the concentrations of Air Quality Index in Haryana Cities. The forecast models were built using air pollutants and meteorological parameters from 2019 to 2021 and testing and validation was conducted on the air quality data for the year 2022 of Jind and Panipat city in the State of Haryana. Further, performance of hybrid forecast model was enhanced using scalar technique and performance was evaluated using various coefficient metrics and other parameters. First, the important factors affecting air quality are extracted and irregularities from the dataset are removed. Second, for forecasting AQI various approaches have been used and evaluation is carried out using performance metrics. The experimental results showed that the proposed hybrid model had a better forecast result than the standard Random forest Regression, Support Vector Regression and Multiple Linear Regression.

A Modeling of Urban Heat Island using Artificial Neural Network and Remote Sensing in Tehran

Zahra Azizi — 2023-11-28

The Urban Heat Island phenomenon happens due to the differences in the thermal behavior between urban and rural areas which many factors such as vegetated, water, impervious and built-up areas could affect this phenomenon. Urban Heat Island consists of three types: Canopy heat island, Boundary heat island, and surface heat island. In this study, the surface type of urban heat island is analyzed. In this paper, 13 TM/ETM+ images have been obtained from 1990 to 2015(an image biennially). Urban Heat Islands effects are much more severe in summer; therefore, all images have been taken in summer. NDVI, IBI, albedo, and also land surface temperature were derived from images. Various neural network topologies have been used to identify the best model for predicting the urban heat island intensity. The LST of 2016 has been considered as validation data, thus the best result from fitting structures was obtained from Cascade which the Bayesian Regularization was its training algorithm (R-squared=0.62, RMSE=1.839 K).

Vegetation Health and Forest Canopy Density Monitoring in The Sundarban Region Using Remote Sensing and GIS

Soma Mitra — 2023-11-28

The present study explores vegetation health and forest canopy density in the Sundarbans region using Landsat-8 images. This work analyzes changes in vegetation health using two vegetation indices, the Normalized Difference Vegetation Index (NDVI) and Forest Canopy Density (FCD) values of the Sundarbans, from 2014 to 2020. NDVI, comprising two bands, Red and Near-infrared (NIR), shows a declining trend during the period. Two NDVI land cover classification maps for 2014 and 2020 are produced, and the interest area is divided into five classes: Scanty, Low, Medium, and Densely Vegetated Regions and Water Bodies. A single-band linear gradient pseudo-color is used to assess the land cover difference between 2020 and 2014, showing marked changes in densely vegetative areas. The NDVI difference marks the coastal regions with a higher depletion rate of vegetation than the regions away from the seacoasts. FCD has been taken to compare the results of NDVI with it. FCD consists of another four models: AVI (advanced vegetative index), BI (Bare soil index), SSI (scaled shadow index), and TI (thermal index). FCD is also called crown cover or canopy coverage, which refers to the portion of an area in the field covered by the crown of trees. 2014 and 2015 FCD maps are produced with a single band linear gradient pseudocolor with five land cover classifications: bare soil, Bare Soil, Shrubs, Low, Medium, and Highly vegetated regions. Both maps bear a significant resemblance to NDVI land classification maps. Further, the FCD values of the two maps are scaled between 1 and 100, and the area of each class is calculated. To check the veracity of the NDVI and FCD analysis, a Deep Neural Network (DNN) model has been developed to classify each year’s image taken from Google Earth Engine (GEE). It classifies each year’s image with 99% accuracy. The calculation of the area of each class emphasizes the rapid decline of densely wooded vegetation. Almost 80% of the highly forested zone has been diminished and has become part of the medium-forested region. Area inflation in medium-forested regions corroborates the same. The study also analyzes the migration of vegetation density, i.e., where and how many areas are unchanged, growing, or deforested.