BMBabita Majhi

Babita Majhi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Babita Majhi has a total of 44 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • JoVE Video Dataset

    by Manishankar Sahu, Babita Majhi, Sujata Dash

    Precision agriculture requires accurate discrimination between winter crops and weeds, but there is a lack of annotated image data for winter cropping systems. This paper investigates a two-stage deep learning approach that integrates self-supervised feature learning with supervised fine-tuning for winter crop and weed image classification. A new winter crop and weed image dataset, WinterCropWeedDB, is proposed and used in this paper, which contains 1,136 high-resolution images of six winter crop species and four weed species collected from agricultural fields in central India. In the first stage of self-supervised learning, an EfficientNet-B3 model is pre-trained using a SimCLR-style self-supervised learning approach with an InfoNCE loss function (temperature τ = 0.5) on the images. The average contrastive loss value reduces from 2.0712 in the first iteration to 1.6835 at the end of pretraining. In the second stage of supervised fine-tuning, the pre-trained EfficientNet-B3 model is fine-tuned with a supervised classifier head on the images and tested on a single internal validation split (30%) of the dataset. The fine-tuned model reaches a maximum validation accuracy of 98.27%, with a macro-averaged F1-score of 0.98. Gradient-weighted class activation mapping (Grad-CAM) and Grad-CAM++ are used on the fine-tuned model to provide a qualitative visualization of the image regions that contribute to class predictions. The experiment outcomes demonstrate the viability of using self-supervised pretraining and supervised fine-tuning for the classification of winter crop and weed images on a region-specific dataset, while also emphasizing the importance of additional testing on independent test sets.

  • A Two-Stage Self-Supervised Learning Framework for Winter Crop-Weed Image Classification

    by Manishankar Sahu, Babita Majhi, Sujata Dash

    Precision agriculture requires accurate discrimination between winter crops and weeds, but there is a lack of annotated image data for winter cropping systems. This paper investigates a two-stage deep learning approach that integrates self-supervised feature learning with supervised fine-tuning for winter crop and weed image classification. A new winter crop and weed image dataset, WinterCropWeedDB, is proposed and used in this paper, which contains 1,136 high-resolution images of six winter crop species and four weed species collected from agricultural fields in central India. In the first stage of self-supervised learning, an EfficientNet-B3 model is pre-trained using a SimCLR-style self-supervised learning approach with an InfoNCE loss function (temperature τ = 0.5) on the images. The average contrastive loss value reduces from 2.0712 in the first iteration to 1.6835 at the end of pretraining. In the second stage of supervised fine-tuning, the pre-trained EfficientNet-B3 model is fine-tuned with a supervised classifier head on the images and tested on a single internal validation split (30%) of the dataset. The fine-tuned model reaches a maximum validation accuracy of 98.27%, with a macro-averaged F1-score of 0.98. Gradient-weighted class activation mapping (Grad-CAM) and Grad-CAM++ are used on the fine-tuned model to provide a qualitative visualization of the image regions that contribute to class predictions. The experiment outcomes demonstrate the viability of using self-supervised pretraining and supervised fine-tuning for the classification of winter crop and weed images on a region-specific dataset, while also emphasizing the importance of additional testing on independent test sets.

  • Open Access

    Prediction of pan evaporation in Chhattisgarh using machine learning techniques

    by Rupesh Naik, Babita Majhi, Diwakar Naidu

    Accurate measurement or estimation of evaporation loss is crucial for developing and successfully implementing water resource management strategies, irrigation planning, reservoir management etc. To predict the pan evaporation (EP) accurately for Raipur, Jagdalpur, and Ambikapur stations of Chhattisgarh, four deep learning models and three machine learning models were used and a hybrid model using Deep Neural Network (DNN) and Random Forest (RF) was proposed. Simulation results demonstrated that the hybrid model (DNN+RF) outperforms the rest with R2 of 0.964, 0.920, 0.894 for Raipur, Jagdalpur and Ambikapur respectively. It has been observed that the hybrid DNN+RF model demonstrated faster convergence compared to other models with high accuracy, making it efficient and well-suited for real-time applications such as irrigation scheduling and water resource management.


  • Open Access

    Forecasting of Daily Pan Evaporation Rate using Deep Learning Techniques for Three Different Agro-Climatic Regions of Chhattisgarh State

    by Babita Majhi, Rupesh Naik, Sujata Dash, Saurav Mallik, Amal Al-Rasheed, Mohamed Abbas, Ben Othman Soufiene

    Accurate measurement or computation of evaporation loss is crucial for developing and successfully implementing water resource management strategies, irrigation planning, reservoir management, hydropower generation, drought and flood mitigation, urban planning and increasing agricultural productivity, especially in drought-prone areas. Evaporation can be measured directly using evaporimeters or forecasted using empirical models based on climatic variables such as temperature, humidity, wind speed, sunlight, and solar radiation, that influence the evaporation process. Modeling evaporation using climatic factors is difficult, especially when accounting for the wide range of agro-climatic conditions as it is an exceedingly nonlinear process. This paper uses different machine learning (ML) and deep learning algorithms to estimate pan evaporation (EP) for three distinct agro-climatic zones in the Indian state of Chhattisgarh. In this research, the performance of three machine learning models (Support Vector Machine, AdaBoost, and XGBoost) and four deep learning models (Deep Neural Network, Recurrent Neural Network, Long Short-Term Memory, and Bidirectional Long Short Term Memory) are evaluated and outcomes from each location are compared. Simulation results demonstrated that across all three regions, deep-Learning models outperform machine-learning and conventional models. Out of all deep learning models DRNN perform the best. As the results exhibit that the (EP) loss per day is less than 1 mm, the proposed model can be used for irrigation scheduling, water resource management which is very important for agriculture and its related activities.

  • Open Access

    A Modified Artificial Neural Network (ANN)-Based Time Series Prediction of COVID-19 Cases from Multi-Country Data

    by Babita Majhi

    The deadly Corona virus that first appeared in a seafood market in the Wuhan city of China in December 2019 has been causing global distress by claiming lives and collapsing economies. Given its serious nature, there is an urgent need to understand the virus’s future trajectory. The current study predicts the next day confirmed, death and recovery cases of COVID-19 pandemic for India, Italy, Spain, and the USA by using a modified multilayer neural network (MMLNN) model. The spread of the COVID-19 data is collected from the Kaggle website for the period of 22nd January 2020 to 20th April 2020 (i.e., for 90 days). The predicted figures of the spread of the disease have been estimated and compared with the actual values. Higher precision of the estimates has been observed from the MMLNN model compared to the conventional multilayer neural network (MLANN) model. Specifically, the MMLNN model does faster and more efficient training of the data resulting in less error. The paper forecasts the next day figures (i.e., for 21st April) for all the three cases and does the comparison of the results with the actual values reported. A deviation of 6% is obtained for India, and for the other three countries the deviation is below 3.5%. Given the high accuracy predictive power, the authors recommend that the MMLNN model can be integrated into the health policy of the countries that are struggling with the spread of the virus. Specifically, a decision on health policies such as restrictions on movement can be based on the short-range predictions of the spread of the virus infection.


  • Open Access

    Estimating the impact of news on Indian government decisions to contain the spread of COVID-19 in India

    by Vikas Bhatnagar, Babita Majhi, Ritanjali Majhi

    Corona-Virus or COVID-19 has adversely affected human life. A large number of human deaths and infected cases were recorded. India has also severely affected an

  • Open Access

    Estimating Long-Run Relationship between Renewable Energy Use and CO2 Emissions: A Radial Basis Function Neural Network (RBFNN) Approach

    by Pradyot Ranjan Jena, Babita Majhi, Ritanjali Majhi

    The long-run relationship between economic growth and environmental quality has been estimated within the framework of the environmental Kuznets Curve (EKC). Several studies have estimated this relationship by using statistical models such as panel regression and time series regression. The current study argues that there is a nonlinear relationship between environmental quality indicators and economic and non-economic predictors and hence an appropriate nonlinear model is required to predict it. An adaptive and nonlinear model, namely radial basis function neural network (RBFNN) has been developed in this study. CO2 emission is used as the target output and renewable energy consumption share, real GDP, trade openness, urban population ratio, and democracy index are used as the predictors to estimate the EKC relationship for nineteen major CO2 emitting countries that account for 78% of the global emissions. The model developed in this study could predict the CO2 emissions of all the countries with more than 95% accuracy. This finding underlines the usefulness of the RBFNN model which can be used to predict emission levels of other pollution indicators at the global level. Further, comparing two models, one with all the predictors and the other excluding the renewable energy share, it was found that the model with renewable energy share predicts CO2 emissions more accurately. This reinforces the already strengthening campaign to encourage industries and governments to increase the share of renewable energy in total energy use.