Uma N. Dulhare has a total of 42 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Enhanced Skin Lesion Classification Using Deep Learning, Integrating with Sequential Data Analysis: A Multiclass Approach

    by Azmath Mubeen, Uma N. Dulhare

    In dermatological research, accurately identifying different types of skin lesions, such as nodules, is essential for early diagnosis and effective treatment. This study introduces a novel method for classifying skin lesions, including nodules, by combining a unified attention (UA) network with deep convolutional neural networks (DCNNs) for feature extraction. The UA network processes sequential data, such as patient histories, while long short-term memory (LSTM) networks track nodule progression. Additionally, Markov random fields (MRFs) enhance pattern recognition. The integrated system classifies lesions and evaluates whether they are responding to treatment or worsening, achieving 93% accuracy in distinguishing nodules, melanoma, and basal cell carcinoma. This system outperforms existing methods in precision and sensitivity, offering advancements in dermatological diagnostics.


  • Open Access

    Understanding and Analysing Causal Relations through Modelling using Causal Machine Learning

    by D. Naga Jyothi, Uma N. Dulhare

    The study of causal inference has gained significant attention in artificial intelligence (AI) and machine learning (ML), particularly in areas such as explainability, automated diagnostics, reinforcement learning, and transfer learning.. This research applies causal inference techniques to analyze student placement data, aiming to establish cause-and-effect relationships rather than mere correlations. Using the DoWhy Python library, the study follows a structured four-step approach—Modeling, Identification, Estimation, and Refutation—and introduces a novel 3D framework (Data Correlation, Causal Discovery, and Domain Knowledge) to enhance causal modeling reliability. Causal discovery algorithms, including Peter Clark (PC), Greedy Equivalence Search (GES), and Linear Non-Gaussian Acyclic Model (LiNGAM), are applied to construct and validate a robust causal model. Results indicate that internships (0.155) and academic branch selection (0.148) are the most influential factors in student placements, while CGPA (0.042), projects (0.035), and employability skills (0.016) have moderate effects, and extracurricular activities (0.004) and MOOCs courses (0.012) exhibit minimal impact. This research underscores the significance of causal reasoning in higher education analytics and highlights the effectiveness of causal ML techniques in real-world decision-making. Future work may explore larger datasets, integrate additional educational variables, and extend this approach to other academic disciplines for broader applicability.


  • Open Access

    Advanced Cotton Boll Segmentation, Detection, and Counting Using Multi-Level Thresholding Optimized with an Anchor-Free Compact Central Attention Network Model

    by Arathi Bairi, Uma N. Dulhare

    Nowadays, cotton boll detection techniques are becoming essential for weaving and textile industries based on the production of cotton. There are limited techniques developed to segment, detect, and count cotton bolls precisely. This analysis identified several limitations and issues with these techniques, including their complex structure, low performance, time complexity, poor quality data, and so on. A proposed technique was developed to overcome these issues and enhance the performance of the detection and counting of cotton bolls. Initially, data were gathered from the dataset, and a pre-processing stage was performed to enhance image quality. An adaptive Gaussian–Wiener filter (AGWF) was utilized to remove noise from the acquired images. Then, an improved Harris Hawks arithmetic optimization algorithm (IH2AOA) was used for segmentation. Finally, an anchor-free compact central attention cotton boll detection network (A-frC2AcbdN) was utilized for cotton boll detection and counting. The proposed technique utilized an annotated dataset extracted from weakly supervised cotton boll detection and counting, aiming to enhance the accuracy and efficiency in identifying and quantifying cotton bolls in the agricultural domain. The accuracy of the proposed technique was 94%, which is higher than that of other related techniques. Similarly, the precision, recall, F1-score, and specificity of the proposed technique were 93.8%, 92.99%, 93.48%, and 92.99%, respectively.


  • Open Access

    Feature Extraction and Identification of Rheumatoid Nodules Using Advanced Image Processing Techniques

    by Azmath Mubeen, Uma N. Dulhare

    Background/Objectives: Accurate detection and classification of nodules in medical images, particularly rheumatoid nodules, are critical due to the varying nature of these nodules, where their specific type is often unknown before analysis. This study addresses the challenges of multi-class prediction in nodule detection, with a specific focus on rheumatoid nodules, by employing a comprehensive approach to feature extraction and classification. We utilized a diverse dataset of nodules, including rheumatoid nodules sourced from the DermNet dataset and local rheumatologists. Method: This study integrates 62 features, combining traditional image characteristics with advanced graph-based features derived from a superpixel graph constructed through Delaunay triangulation. The key steps include image preprocessing with anisotropic diffusion and Retinex enhancement, superpixel segmentation using SLIC, and graph-based feature extraction. Texture analysis was performed using Gray-Level Co-occurrence Matrix (GLCM) metrics, while shape analysis was conducted with Fourier descriptors. Vascular pattern recognition, crucial for identifying rheumatoid nodules, was enhanced using the Frangi filter. A Hybrid CNN–Transformer model was employed for feature fusion, and feature selection and hyperparameter tuning were optimized using Gray Wolf Optimization (GWO) and Particle Swarm Optimization (PSO). Feature importance was assessed using SHAP values. Results: The proposed methodology achieved an accuracy of 85%, with a precision of 0.85, a recall of 0.89, and an F1 measure of 0.87, demonstrating the effectiveness of the approach in detecting and classifying rheumatoid nodules in both binary and multi-class classification scenarios. Conclusions: This study presents a robust tool for the detection and classification of nodules, particularly rheumatoid nodules, in medical imaging, offering significant potential for improving diagnostic accuracy and aiding in the early identification of rheumatoid conditions.


  • AI-Driven Personalization in Tourism

    by A. V. Senthil Kumar, M. Mosikan, Amit Dutta, Manjunatha Rao L., Ismail Bin Musirin, Venkata Shesha Giridhar Akula, Saifullah Khalid, Suresh Babu Yalavarthi, Chandra Shekar D. V., Suganya R. V., et al.

    This chapter delves into the transformative impact of artificial intelligence (AI) technologies on the tourism sector, focusing on personalized and sustainable travel experiences. It explores how AI-driven recommendation systems, chatbots, and predictive analytics revolutionize personalization by analyzing vast datasets to tailor travel recommendations based on individual preferences. The chapter also addresses challenges such as data privacy and ethical considerations, emphasizing the need for responsible AI implementation. Looking ahead, the chapter underscores the potential for further AI-driven innovations to enhance personalized and sustainable tourism experiences, paving the way for a more efficient, enjoyable, and eco-conscious travel industry. Through AI, the tourism industry can offer travellers highly personalized experiences while minimizing their environmental impact. As AI technologies continue to advance, there is a growing opportunity to create a more sustainable and responsible tourism sector that meets the needs of both travellers and the planet.


  • Advancements in Metaverse Security

    by A. V. Senthil Kumar, Pavithra Sivakumar, Ankita Chaturvedi, Ismail Bin Musirin, Venkata Shesha Giridhar Akula, R. V. Suganya, G. Vanishree, Rajani H. Pillai, G. Jagadamba, Gaganpreet Kaur, et al.

    This chapter proposes a novel approach for detecting phishing websites within the metaverse, leveraging the Optimal Feature Selection and the Random Forest classifier. This framework addresses the critical challenge of safeguarding users from deceptive tactics in virtual environments. By analyzing website characteristics and identifying the most informative features, the proposed method enhances the accuracy and efficiency of phishing detection in the metaverse, contributing to a more secure and trustworthy virtual landscape. The chapter delves into the methodology, including the chosen feature selection technique and the Random Forest classifier, followed by implementation details, experimental results evaluating the model's performance, and a discussion on the implications for future metaverse security research


  • Optical Character Recognition (OCR) Using Opencv and Python

    by Ajay Karthick M. A. V. Senthil Kumar, Ajay Karthick M., Ahmad Fuad Hamadah Bader, Gaganpreet Kaur, Samrat Ray, Prasanna Lakshmi G., Paresh Virparia, Bharat Bhushan Sagar, Amit Dutta, Shadi R Masadeh, et al.

    Optical character recognition (OCR) stands as a transformative technology at the intersection of computer vision and document processing. This chapter explores the advancements and challenges in OCR, focusing on methods for extracting text content from images, scanned documents, and other visual media. The review encompasses traditional techniques, such as template matching and feature-based methods, as well as state-of-the-art deep learning approaches. The evolution of OCR algorithms is discussed in the context of their applications in digitizing historical archives, automating data entry, enhancing accessibility, and facilitating language translation. Additionally, attention is given to challenges related to diverse fonts, handwriting recognition, and handling complex document layouts. The chapter concludes with an outlook on emerging trends and future directions in OCR research, emphasizing the ongoing pursuit of accuracy, robustness, and efficiency in extracting textual information from visual data.


  • Agricultural Crop Recommendations Based on Productivity and Season

    by Aparna M. A. V. Senthil Kumar, Aparna M., Amit Dutta, Samrat Ray, Hakikur Rahman, Shadi R. Masadeh, Ismail Bin Musirin, Manjunatha Rao L., Suganya R. V., Ravisankar Malladi, et al.

    This chapter aims to develop an agricultural crop recommendation system leveraging the power of machine learning algorithms. The proposed system takes into account crop productivity and prevailing season as crucial factors in making appropriate crop suggestions. The authors proposed the SVM algorithm, which was trained and evaluated on a comprehensive dataset comprising historical agricultural data with diverse features such as climate variables, soil properties, and geographical factors. The data was further segmented based on seasonal patterns to provide crop recommendations tailored to specific timeframes. The models' performance was evaluated using standard metrics, and an ensemble approach was considered to enhance the system's robustness. Ultimately, the developed system offers farmers and agricultural experts a valuable tool for making informed decisions, optimizing crop selection, and increasing overall agricultural productivity