Anusha Papasani
@vitap.ac.in
Ph.D Research scholor
VIT-AP University
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Anusha Papasani, , Nagaraju Devarakonda, and
Elsevier BV
An important task for classification is feature selection that removes the redundant or irrelevant features from the dataset. Multi-objective feature selection approach is mainly proposed by many researchers. However, these approaches failed to maintain the higher classification accuracy while removing redundancy in the features. In this work, a wrapper based feature selection technique is proposed with a hybrid of Multi Objective Honey Badger Algorithm (MO-HBA) and Strength Pareto Evolutionary Algorithm-II to maintain the balance between classification accuracy and removal of redundancy. Classification accuracy improvement and removal of redundant features are considered as the multi-objective optimization functions of the proposed multi-objective feature selection technique. The Levy flight algorithm is utilized to initialize the population to enhance the ability of the exploration and exploitation of MO-HBA. The regularized Extreme Learning Machine is used to classify the selected features. To evaluate the performance of the proposed feature selection technique, eighteen benchmark datasets are utilized and results are compared with the four well known multi-objective feature selection techniques in terms of accuracy, hamming loss, ranking loss, mean value, standard deviation, length of features, and training time. The proposed approach achieved maximum accuracy of 100% with the maximum value of selected features as 80. The minimum value of hamming loss, ranking loss, mean value and standard deviation value achieved by the proposed approach are 0.0092, 0.0003, 0.018 and 0.001 respectively. The experimental results show that the proposed approach can give improved classification accuracy while the removal of redundancy in large scale datasets.
Anusha Papasani and Nagaraju Devarakonda
Elsevier BV
Anveshini Dumala, Anusha Papasani, Rajeswari Bommala, and Vikkurty Sireesha
IEEE
Detection of plant disease at an early stage increases the crop yield otherwise these diseases may negatively impact the agro market economy. The conventional methods were time consuming and practically infeasible to cover thousands acres of farming areas to detect leaf diseases. A methodology is proposed in this paper, to spot and to analyze the plant leaf diseases using digital image processing techniques through a supervised machine learning technique called multi-support vector machine (m-SVM) algorithm. SVM handles both semi structured and unstructured data. The proposed model recognizes and classifies the images of the leaves that were captured by digital camera or a mobile phone or drones or web camera. A novel way of training and methodology was used to accelerate the speedy, easy and simple implementation of the system in real-time. The experimental outcomes make evident that the proposed system detects and classifies the major 6 plant leaves diseases successfully: Cercospora leaf spot, Alternaria Alternata, Rust, Anthracnose, Powdery Mildew and Bacterial Blight. Also some of the unanswered challenges are discussed that require to be answered by developing a sensible automatic plant disease recognition system to apply in field conditions.
Anusha Papasani, Nagaraju Devarakonda, and Zdzislaw Polkowski
IEEE
Anveshini Dumala, Anusha Papasani, and Sireesha Vikkurty
Springer Singapore
Severe acute respiratory syndrome coronavirus (SARS-CoV) is recognized, and very first person infected is from the Guangdong province of southern China in 2002 while the virus that causes COVID-19 (Corona VIrus Disease-2019) is known as SARS-CoV-2. World Health Organization (WHO) named it as “COVID-19” on February 11, 2020. Currently, the COVID-19 has frightened the whole world of human beings and pushed into the pandemic. This coronavirus affects the respiratory system by entering into the human body through the droplets of saliva and mucus. It takes 14 days to observe the symptoms of the virus attack. In the meantime, the virus affected person may spread the virus to the coexisting people in the abode unknowingly. Also, it takes 48 h to confirm if a person is virus attacked after the test sample is collected. So, there is a serious need to wear a face mask that covers the nose and mouth besides maintaining the social distance to break the chain of massive increase. This paper attempts to detect if an individual wears a mask, using OpenCV. The accurate identification of landmarks of face in the image is an imperative challenge. Being instinctive it is simple for a human to detect the object, but it took years of research to raise the accessibility of quality datasets and a remarkable progress. The purpose of the paper is identifying the count of faces with the mask in the image and count of faces without a mask on live webcam. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
B. Mounika, P. Anusha, V. Narayana and G. V. Lakshmi
SynthesisHub Advance Scientific Research
This article overviews a chain based methodologies for a few security administrations. The Existing framework is contrasted and the proposed framework and it was discovered that the proposed framework has preferred execution over the existing one. Square chain offers a creative way to deal with putting away data, executing exchanges, performing capacities, and building up trust in an open situation. Many consider the square chain as an innovative leap forward for cryptography and digital security. Square chain administrations incorporate verification, secrecy, security and access control list (ACL), information and asset provenance, and uprightness affirmation. Every one of these administrations is basic for the current appropriated applications, particularly because of the huge measure of information being handled over the systems and the utilization of distributed computing. Square chain method is furnished with validation, inspecting, and responsibility, and consequently, it can fill in as a promising instrument for giving secure information correspondence on the system. Validation guarantees that the client is who he/she professes to be. Privacy ensures that information can't be perused by unapproved clients. Protection gives the clients the capacity to control who can get to their information. Provenance permits an effective following of the information and assets alongside their proprietorship and usage over the system. Trustworthiness helps in checking that the information has not been changed or adjusted. These administrations are right now overseen by concentrated controllers, for instance, a declaration authority. Along these lines, the administrations are inclined to assaults on the incorporated controller. Then again, the square chain is made sure about and disseminated records that can help settle a large number of the issues with centralization. From a security viewpoint, the square chain is made and kept up utilizing a distributed overlay arrangement and made sure about through shrewd and decentralized use of cryptography with swarm processing. Block chain offers an imaginative way to deal with putting away data, executing exchanges, performing capacities, and setting up a trust in an open environment. Many consider the square chain as an innovative leap forward for cryptography and digital security, with use cases running from internationally conveyed digital money frameworks. A decentralized distributed storage arranges has been presented with numerous favorable circumstances over the server farm based capacity. Comparable to conventional arrangement, decentralized distributed storage organize use customer side encryption to keep up information security.
V. Narayana, B. Sudheer, Venkata Rao Maddumala and P. Anusha
SynthesisHub Advance Scientific Research
Content Extraction assumes a significant job in discovering essential and important data. Content extraction includes discovery, restriction, following, binarization, extraction, improvement and acknowledgment of the content from the given picture. This paper proposes a bi-leveled picture characterization framework to group printed and transcribed reports into totally unrelated predefined classes. In the present article, we propose a Fuzzy guideline guided novel strategy that is utilitarian without any outer intercession during execution. The Fuzzy validation processor utilizes subtleties of divider speed and force, just as change, to decide if the deliberate reverberation signal part to be separated genuinely speaks to the divider speed as it were. Test results propose that this methodology is a proficient one in contrast with various different strategies widely tended to in writing. At long last, the exhibition of the proposed framework is contrasted and the current frameworks and it is seen that proposed framework performs superior to numerous different frameworks.