Search

The classification of land cover objects in hyperspectral imagery (HSI) has significantly advanced due to the development of convolutional neural networks (CNNs). However, challenges such as limited training data and high dimensionality negatively impact classification performance. Traditional CNN-based methods predominantly utilize 2D CNNs for feature extraction, which inadequately exploit the inter-band correlations in HSIs. While 3D CNNs can capture joint spectral-spatial information, they often encounter issues related to network depth and complexity. To address these issues, we propose an innovative land cover object classification approach in HSIs that integrates segmented principal component analysis (Seg-PCA) with hybrid 3D-2D CNNs. Our approach leverages Seg-PCA for effective feature extraction and employs the minimum-redundancy maximum relevance (mRMR) criterion for feature selection. By combining the strengths of both 3D and 2D CNNs, our method efficiently extracts spectral-spatial features. These features are then processed through fully connected dense layers and a softmax layer for classification. Extensive experiments on three widely used HSI datasets demonstrate that our method consistently outperforms existing state-of-the-art techniques in classification performance. These results highlight the efficacy of our approach and its potential to significantly enhance the classification of land cover objects in hyperspectral imagery. [ABSTRACT FROM AUTHOR]

We present an innovative hyperspectral image (HSI) classification method addressing challenges posed by closely spaced wavelength bands. Our approach combines 3D-2D convolutional neural networks (CNNs) with multi-branch feature fusion for improved spectral-spatial feature extraction. Using segmented principal component analysis (Seg-PCA), we reduce HSIs' spectral dimensions into global and local intrinsic characteristics. The integration of 3D and 2D CNNs captures joint spectral-spatial features, while a multi-branch network extracts and merges diverse local features along the spectral dimension. Our method outperforms existing approaches, achieving remarkable accuracy of 99.27%, 100%, and 99.99% on Indian Pines, Salinas Scene, and University of Pavia datasets, respectively. [ABSTRACT FROM AUTHOR]

Farm programmes (FPs) of varied categories have been developed and aired over several decades by Bangladesh Betar, the national radio of Bangladesh for the diffusion of farm technologies. The study aimed to produce an in-depth academic evaluation of their effectiveness in educating farmers in Bangladesh. A sample of 465 respondents from the Khulna and Rajshahi divisions in Bangladesh was randomly selected for a questionnaire survey. To analyse the data, relevant documents were collected from the Ministry of Information, the Ministry of Agriculture, and the Ministry of Fisheries and Livestock of Bangladesh. Frequency distribution, z-test, and binary logistic regression analysis were used as statistical tools. The farm knowledge levels of the farmers were considered the predictors for evaluating the effectiveness of FPs. The results revealed that 93.33% of the sample did not listen to the FPs, while only very few of the listeners listened regularly. Despite this, at the 'weak' and 'average' levels of knowledge significant differences were noted between listener and non-listener farmers of FPs. The binary logistic regression analysis (Model 1) identified that the farmers who listened to the FPs were likely to acquire farm knowledge 6.62 times more than the farmers who did not listen to the FPs. The farmers who listened to the FPs were likely to have farm knowledge 2.64 times more than the farmers who did not listen to the FPs but consulted with other sources of farming information (Model 2). Similarly, a listener of FPs with farm training was likely to acquire farm knowledge 5.76 times more than a non-listener with farm training (Model 3). The FPs were found to be very effective and could be used to better complement other mechanisms for educating farmers. Regular access to the FPs ought to be ensured through appropriate stimulants for the diffusion of farm technologies.

The use of hyperspectral imaging sensors has greatly improved the classification of remotely sensed data because of the abundant spectral information they offer. However, the numerous contiguous, tiny wavelength bands captured in hyperspectral images (HSIs) often hinder the classification process. To overcome the aforementioned problem, various feature reduction techniques, including feature extraction (FE) and feature selection (FS), are commonly employed to improve classification performance. Linear discriminant analysis (LDA) is a well-established approach that has been utilized for the FE of HSI. LDA's consideration of global characteristics and the variance accumulator for FS can lead to the poor reduction of HIS's intrinsic characteristics. Furthermore, LDA's limited ability to select a very low number of features, i.e. the number of classes minus one, restricts its effectiveness in HSI classification. Therefore, we introduce an FS method based on a non-linear information-theoretic measure, normalized mutual information (nMI) combined with minimum redundancy maximum relevance (mRMR). This approach is used to identify inherent features from the transformed space of our proposed correlation-based segmented-LDA (SLDA) and spectral region-based segmented-LDA (SSLDA) FE methods. We thoroughly compare the performance of the SLDA-mRMR and SSLDA-mRMR methods with the existing linear and non-linear state-of-the-art techniques, including unsupervised principal component analysis (PCA), PCA-based methods, supervised LDA, and LDA-based methods. Additionally, we explore the performance of nMI-based mRMR selection with all FE methods by incorporating a cumulative variance-based top features pick-up strategy. Based on the experimental results, we observe that SSLDA-mRMR and SLDA-mRMR achieve the highest classification accuracy, such as 91.91%, and 91.57% for agricultural Indian Pines, respectively, 97.66%, and 97.54%, respectively, for Kennedy Space Center, and 96.69%, and 96.57%, respectively, for Pavia University. In contrast, the classification accuracies using all original features of the HSIs are 71.39%, 70.01%, and 83.52%, respectively. Moreover, our proposed SSLDA-mRMR and SLDA-mRMR combinations outperformed all other FE and FS combinations that were examined using actual HSI datasets. [ABSTRACT FROM AUTHOR]

Background: Healthy life expectancy (HALE) at birth measures the standard of living anticipated. Objective: The purpose of this study was to identify the main global risk factors for HALE at birth. Methodology: The data of 212 countries have been extracted from the data hub of World Health Organization, Worldometer, World Bank, and United Nations. The HALE at birth was considered as the dependent variable and the death rate due to Coronavirus disease 2019 (COVID 19), recovery rate from COVID 19, tuberculosis (TB) incidence, universal health coverage (UHC) service index, tobacco smoking, alcohol consumption rate, human immunodeficiency virus (HIV) prevalence rate, average household size, gross domestic product (GDP), and current health expenditure were considered as the independent variables. Descriptive statistics, Pearson's correlation analysis, and multiple linear regression model were used as the statistical tools to analyze the data. Results: The study results revealed that the HALE is found lowest in Central African Republic (44.9 years) and highest in Singapore (76.2 years). On the other hand, the death rate due to COVID 19 is found highest in Nicaragua (32.0%). In case of the recovery rate from COVID 19, the lowest value is seen in Tajikistan (0.0). Again, the lowest UHC service index is found in Montserrat (0.0). Alcohol consumption rate is found highest in Moldova Republic (15.20). In case of HIV prevalence rate, the highest value is seen in Eswatini (27.30). Average household size is seen highest in Senegal (8.30). The recovery rate from COVID 19, UHC service index, GDP, current health expenditure TB incidence, tobacco smoking, HIV prevalence rate and average household size were significantly correlated with the HALE at birth. Again, the multiple linear regression model identified the UHC service index, alcohol consumption rate, HIV prevalence rate and average household size as the significant factors which are responsible for the variation in HALE at birth worldwide. Conclusion: The UHC service index, alcohol consumption rate, HIV prevalence rate and average household size are the most associate factors of HALE at birth globally. [ABSTRACT FROM AUTHOR]

Tackling malnutrition is a major health priority for a developing country like Bangladesh. This study explored the differences in prevalence of having only one form, and multiple forms, of severe malnutrition (stunting, wasting and underweight) among under-5 children in Bangladesh, and aimed to identify the important factors affecting these. Data were extracted from the Bangladesh Demographic Health Surveys conducted in 2007, 2011 and 2014. The outcome measures were 'only one form' and 'multiple forms' of severe malnutrition in children aged under 5 years. A Chi-squared test was performed to find the association of outcome variables with selected socio-demographic factors and logistic regression models were applied to identify risk factors. A total of 19,874 children aged under 5 years were included in the analysis. The overall proportion with one form of severe child malnutrition was approximately 12%, and the proportion with multiple forms was 8%. Age, mother's education, father's occupation, mother currently working, watching television, source of water, solid waste used in cooking, intimate partner violence (IPV), wealth index, urban/rural place of residence and birth cohort were found to be significant factors for both having only one and having multiple forms of severe child malnutrition. Children with an uneducated mother of poor socioeconomic class had a higher risk of severe malnutrition. Children of fathers with a professional occupation were at lower risk of having multiple forms of severe malnutrition. The proportions of children aged under 5 years with one or multiple forms of severe malnutrition were shown to be high in Bangladesh. The prevention of malnutrition in the country should be seen as a significant public health issue and given top priority. [ABSTRACT FROM AUTHOR]