PSR-LeafNet: A Deep Learning Framework for Identifying Medicinal Plant Leaves Using Support Vector Machines Praveen Kumar Sekharamantry, Marada Srinivasa Rao, Yarramalle Srinivas, Archana Uriti Big Data and Cognitive Computing, 2024 In computer vision, recognizing plant pictures has emerged as a multidisciplinary area of interest. In the last several years, much research has been conducted to determine the type of plant in each image automatically. The challenges in identifying the medicinal plants are due to the changes in the effects of image light, stance, and orientation. Further, it is difficult to identify the medicinal plants due to factors like variations in leaf shape with age and changing leaf color in response to varying weather conditions. The proposed work uses machine learning techniques and deep neural networks to choose appropriate leaf features to determine if the leaf is a medicinal or non-medicinal plant. This study presents a neural network design based on PSR-LeafNet (PSR-LN). PSR-LeafNet is a single network that combines the P-Net, S-Net, and R-Net, all intended for leaf feature extraction using the minimum redundancy maximum relevance (MRMR) approach. The PSR-LN helps obtain the shape features, color features, venation of the leaf, and textural features. A support vector machine (SVM) is applied to the output achieved from the PSR network, which helps classify the name of the plant. The model design is named PSR-LN-SVM. The advantage of the designed model is that it suits more considerable dataset processing and provides better results than traditional neural network models. The methodology utilized in the work achieves an accuracy of 97.12% for the MalayaKew dataset, 98.10% for the IMP dataset, and 95.88% for the Flavia dataset. The proposed models surpass all the existing models, having an improvement in accuracy. These outcomes demonstrate that the suggested method is successful in accurately recognizing the leaves of medicinal plants, paving the way for more advanced uses in plant taxonomy and medicine.
Evaluating Object Detection Approaches for Fruit Detection in Precision Agriculture: A Comprehensive Review Archana Uriti, Naga Jyothi Pothabathula 2024 International Conference on Intelligent Computing and Sustainable Innovations in Technology IC Sit 2024, 2024 Agricultural automation has become increasingly vital in addressing the growing demand for food and the need for efficient farming practices. Fruit harvesting is crucial in agriculture due to its labor intensity and the challenges of timely, accurate picking under occlusions and varying illumination. Various methods have been developed for accurate fruit detection, but traditional techniques often struggle with occlusions by branches or leaves, as well as challenges posed by climate and fruit maturity. This study reviews existing literature, compares various fruit detection approaches and discusses their strengths and weaknesses. Building on the work of various authors, this paper aims to offer a thorough understanding of the current state of fruit detection technologies in agricultural automation. Additionally, the survey addresses the objective of the proposed work, which is to identify the most promising techniques for improving fruit detection accuracy and efficiency.
A Systematic Analysis of Multi-Property Prediction Using Deep Learning in the Field of Drug Development Archana Uriti, Abhisek Sethy, Surya Prakash Yalla 2024 International Conference on Intelligent Computing and Sustainable Innovations in Technology IC Sit 2024, 2024 In the rapidly changing field of pharmaceutical development, it is crucial to speed up clinical trials in order to introduce effective drugs in a timely manner. Conventional methods of forecasting molecular characteristics are often slow and time consuming, causing difficulties in the drug discovery process. However, Advancements in deep learning have streamlined pharmaceutical testing, enabling precise prediction of key drug properties like solubility, toxicity, reactivity, and bioactivity before synthesis. By employing different Simplified Molecular Input Line Entry System (SMILES) representations and training them with various deep learning algorithms like Graph Neural Networks (GNN) and Graph Convolution Networks (GCN). In this, a comparative study done on various technologies used in predicting molecular properties that helps to develop faster clinical trials in pharmaceutical industry. By identifying the limitations in the existing technologies, proposed an improved GCN which helps in better estimation of molecular properties.
Secure medical sensor monitoring framework using novel optimal encryption algorithm driven by Internet of Things J. Lekha, K. Sandhya, Uriti Archana, Chunduru Anilkumar, Saini Jacob Soman, S. Satheesh Measurement Sensors, 2023 Recently, healthcare monitoring systems have emerged as significant tolls for constant monitoring of patient's physiological characteristics. These systems use implanted sensors. IoT (Internet of Things) have revolutionized healthcare systems where health care equipment's are equipped with many sensors that actively collect data from patients and pass it on to cloud based storages using gateway sensors. Securing data have been significant barriers in many applications as false information get injected, or important information are modified or stolen at different phases of health care systems dependent on IoT. The attacks can also result in fatalities making it imperative to secure IoT based health care systems. A Hybrid technique combining MOAES (Modified Optimal Advanced Encryption Standard) with CM (Chaotic Map) Encryptions called HMOAES-CM technique is proposed. This technique can be helpful in securely accessing the patient data over online mode, and in addition, the data sharing can be performed in an encrypted form for the necessary targets of stakeholders. The proposed authentication approach is aimed at IoT, which is resilient to all kinds of network attacks and its implementation is also simpler. Comparing the suggested work to similar works, the level of evaluation is much improved.
Exploration on Quick Response (QR) Code Behaviour in Commerce based Platforms Using Machine Learning , Archana Uriti, Surya Prakash Yalla International Journal of Information Engineering and Electronic Business, 2023 The "rapid response" code, or QR code, is made to quickly decode vast amounts of data. Any managed device, such as a smartphone, is able to capture it, and it is simple to access simply scanning the 2D matrix code. The dataset is analyzed utilizing machine learning techniques, such as the confusion matrix score utilized for the multinomial naive Bayes algorithm's performance analysis. The QR code generation is limited to single product and is extended now to include all products. Due to its ability to provide clients with benefits including speedy, error-free access and the ability to store a lot of data. Generally, many people are using the online payment for any transaction for flexibility and one can do at any place at any time. For bulk or huge payment, cash is not a good option. Hence many retailers join in the e-wallet companies and make their payment so flexible and faster transaction. Because of these benefits, QR code has becoming widespread.
A comprehensive machine learning framework for automated book genre classifier Abhisek Sethy, Ajit Kumar Rout, Archana Uriti, Surya Prakash Yalla Revue D Intelligence Artificielle, 2023 Machine learning has been leveraged in the digital era, resulting in an increasing desire for computers to perform human-like tasks.Text classification is rapidly becoming one of the most significant applications of machine learning.However, the manual reading and classification of books based on genre requires substantial time and effort.As a result, machine learning methods are critical for enabling automated classification.In this study, a book description-based text classification framework was proposed, utilizing a wealth of information about book contents.The automated classification of books was achieved through the implementation of supervised machine learning.A variety of classifiers were employed, including Multinomial Naive Bayes, Gradient Boosting, and Random Forest, to categorize book genres.According to the results, the Naive Bayes classifier outperformed the other two techniques in classification accuracy, while comparable performance was achieved with Gradient Boosting and Random Forest.The comprehensive machine learning framework efficiently and accurately categorized books by extracting information from book descriptions.The proposed methodology has the potential to facilitate large-scale book classification for both academic and industrial purposes.Overall, this study provided an automated solution to relieve the burden of manual classification while achieving high accuracy.
A Deep Belief Network Based Land Cover Classification Koteswara Rao Chintada, Surya Prakash Yalla, Archana Uriti 3rd IEEE International Virtual Conference on Innovations in Power and Advanced Computing Technologies I Pact 2021, 2021
