BAB. Ankayarkanni

B. Ankayarkanni

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


B. Ankayarkanni has a total of 19 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • A Food Delivery System: Using MERN Stack and Machine Learning

    by Debankana Dhar, Debadrita Dhar, B. Ankayarkanni

  • Open Access

    Enhanced ResNet-50 with Multi-Feature Fusion for Robust Detection of Pneumonia in Chest X-Ray Images

    by Neenu Sebastian, B. Ankayarkanni

    Background/Objectives: Pneumonia is a critical lung infection that demands timely and precise diagnosis, particularly during the evaluation of chest X-ray images. Deep learning is widely used for pneumonia detection but faces challenges such as poor denoising, limited feature diversity, low interpretability, and class imbalance issues. This study aims to develop an optimized ResNet-50 based framework for accurate pneumonia detection. Methods: The proposed approach integrates Multiscale Curvelet Filtering with Directional Denoising (MCF-DD) as a preprocessing step to suppress noise while preserving diagnostic details. Multi-feature fusion is performed by combining deep features extracted from ResNet-50 with handcrafted texture descriptors such as Local Binary Patterns (LBPs), leveraging both semantic and structural information. Precision attention mechanisms are incorporated to enhance interpretability by highlighting diagnostically relevant regions. Results: Validation on the Kaggle chest radiograph dataset demonstrates that the proposed model achieves higher accuracy, sensitivity, specificity, and other performance metrics compared to existing methods. The inclusion of MCF-DD preprocessing, multi-feature fusion, and precision attention contributes to improved robustness and diagnostic reliability. Conclusions: The optimized ResNet-50 framework, enhanced by noise suppression, multi-feature fusion, and attention mechanisms, offers a more accurate and interpretable solution for pneumonia detection from chest X-ray images, addressing key challenges in existing deep learning approaches.


  • Development of a virtual reality system for aquatic ecosystem exploration and education

    by Umesh Daivagna, Ankayarkanni B, Biswaranjan Swain, Nachappa M.N, Shobhit Goyal, Takveer Singh, Mekala Ishwarya

    A computer-generated simulation made up of hardware and software is called virtual reality (VR). Immersive virtual reality (IVR) and non-immersive VR are the two types of VR that have been developing for years. The most recent version gives the impression of physically presenting in a non-physical setting, whereas the previous uses a computerbased environment that displays a simulated reality. The concerns surrounding the usage of a virtual reality application as a tool to expose students to the actual world while they are studying and gaining practical knowledge of a site as it is depicted in the virtual environment are the main focus of this work in e-learning. The topics covered in the study pertain to the creation of virtual reality applications from the standpoint of combining various forms of data about the actual world and educational resources. The study's lessons could be applied to other educational and training settings, such architecture, transportation, and firefighting training—all of which need students to apply what they have learnt in the classroom to real-world scenarios. The findings of this study have a number of significant ramifications for scholars and developers.


  • Exposing users’ emotional surveillance in social networks using machine learning

    by Kurukuri Sai Shankar, Morasa Chandra Surya Prakash, D. Usha Nandini, B. Ankayarkanni, A. Mary Posonia

    Social media is a famous person to person communication site. It is a most common internet based long reach casual correspondence benefit the enable client to s Social media is a famous person to person communication site. It is a most common internet based long reach casual correspondence benefit the enable client to s

  • Open Access

    CART- A Statistical Model for Predicting QoE using Machine Learning in Smartphones

    by Ruth Christina, Greeshma Liz Shajan, B. Ankayarkanni

    The proliferation of fast, cheap, and ubiquitous network access, in particular on mobile devices, has significantly changed the way user’s access online benefits on an everyday premise. Alongside the expanded accessibility of a dependably on Internet, new advancements, for example, LTE have empowered the utilization of a wide range of administrations, with stringent necessities regarding system execution and limit requests. We consider different QoE appraisal models in view of administered machine learning strategies, which are proficient to foresee the QoE experienced by the end client of well-known cell phone applications (e.g., YouTube and Face book), utilizing as info the inactive in-gadget estimations. In this extraordinary issue, we investigate the connection between the Quality of Service (QoS) that administrators screen and oversee and the Quality of Experience (QoE) that the clients really get. Utilizing a rich QoE dataset got from field trials in operational cell systems, we benchmark the execution of various machine learning based indicators, and build a choice tree based model which is able to foresee the per-client general understanding and administration worthiness with a win rate of 91% and 98% separately. To the best of our knowledge, this is the initial proposal using end-user, in-device passive measurements and machine learning models to predict the QoE of smartphone users in operational cellular networks.


  • Open Access

    Performance Analysis of Edge Detection Algorithms on Various Image Types

    by Tamilselvi Nagasankar, B. Ankaryarkanni

    Objective: Various edge detection algorithms are analyzed to find the best and worst performance of edge detection algorithm on various image types. Methods/Statistical Analysis: Only .tif image files are considered for the analysis. Some of the sample images in MATLAB tools and some from web are considered as source for the performance analysis. The performance of the edged image is measured using the entropy and signal noise ratio. High entropy and SNR values specified the high quality of the edged image and the low values indicated the low quality of the image. Findings: Making a deep analysis on various edge detection algorithms is really worth enough in Image processing. Here, five commonly used edge detection algorithms such as Prewitt, Sobel, Robert, Log and Canny are consider for analysis. Matrix form of grayscaled, graysliced, indexed, binary and dither binary image information are taken for the analysis. The analysis is done to find the best and worst performance of edge detection algorithm on various image types. For an image, five different edge detection algorithms applied on five different image information. Totally twenty five edged images are generated as output of an image. From the analysis, it is identified that Canny edge detection algorithm is performing better among the five algorithms. Out of the five image information, Canny algorithm on Dither binary image information yields the high entropy and SNR values. But, the Robert algorithm with indexed image information generates the very low entropy with low SNR values. Applications/Improvements: Edge detection is an important and basic operation to be completed for any image processing activities, image analysis, pattern recognition on various images such as satellite images, medical images etc.,


  • Open Access

    A Description of Content Based Image Retrieval using from Block Truncation Coding and Image Content Description

    by D. Venkatakrishna, B. Ankayarkanni

    Background: The paper provides framework description for survey of content based image recovery framework block truncation coding for image content description. Methods and Statistical Analysis: The technique of content based image recovery framework by exploiting low level complexity image and block truncation coding is one of lossy picture pressure technique for grayscales image. Findings: And here proposes a idea of content based picture recovery utilizing block truncation coding for image content description using Content Based Image Retrieval (CBIR) algorithm and image content description is content-based indexing and recovery of pictures are make necessary semantic portrayal for picture content description of mapping from low-level to abnormal state picture application specifics. Applications/Improvement: content based image recovery framework method improves the performance of the existing method


  • A new statistical rule model for image retrieval system of remote sensing images

    by B Ankayarkanni, A Ezil Sam Leni

    RGIRS (Remote Geo-system Image Retrieval System) is a system of retrieving similar image using image features like color feature, texture feature and shape feature. Content based image retrieval system extracts features relevant to query image using feature extraction method. Many RGIRS systems are proposed to retrieve accurate similar image but the problem is no method provides accurate results. In this paper a new Statistical Rule Model is proposed to retrieve the image from the database using threshold based Euclidean distance calculation. The input query image features are extracted and sent to the SVM Model. The SVM will compare input image features with trained features and test the features with distributed density function. Once the image class is identified, the features are rearranged and calculate the distance using Euclidean distance calculation. A threshold value is fixed for retrieval. If the calculated distance is less than the threshold value the image will be retrieved or if the distance exceeds threshold value then the image will not be retrieved.


  • An efficient image retrieval system for remote sensing images

    by B Ankayarkanni, A Ezil Sam Leni

    RGIRS (Remote Geo-system Image Retrieval System) is a system of retrieving similar image using image features like color feature, texture feature and shape feature. Content based image retrieval system extracts features relevant to query image using feature extraction method. Many RGIRS systems are proposed to retrieve accurate similar image but the problem is no method provides accurate results. In this paper a new Statistical Rule Model is proposed to retrieve the image from the database using threshold based Euclidean distance calculation. The input query image features are extracted and sent to the SVM Model. The SVM will compare input image features with trained features and test the features with distributed density function. Once the image class is identified, the features are rearranged and calculate the distance using Euclidean distance calculation. A threshold value is fixed for retrieval. If the calculated distance is less than the threshold value the image will be retrieved or if the distance exceeds threshold value then the image will not be retrieved.