Research papers

"Research paper." What image comes into mind as you hear those words: working with stacks of articles and books, hunting the "treasure" of others' thoughts? Whatever image you create, it's a sure bet that you're envisioning sources of information--articles, books, people, and artworks. Yet a research paper is more than the sum of your sources, more than a collection of different pieces of information about a topic, and more than a review of the literature in a field. A research paper analyzes a perspective or argues a point. Regardless of the type of research paper you are writing, your finished research paper should present your own thinking backed up by others' ideas and information.

Here are some publications by AAC students.

Title:

Smart Posture Detection and Correction system using Skeletal Points Extraction

Abstract:

This paper is intended to present a smart posture recognition and correction system. In specific, sitting in wrong posture for persistent period of time results in many health problems such as back pain, soreness, poor circulation, cervical pains and also decrease in eyesight in the long run. The proposed model makes use of real time skeletal points extraction. This system is based on computer vision and machine learning algorithms.

Authors:

Publication:

Advances in Decision Sciences, Image Processing, Security and Computer Vision (pp.177-181), January 2020,
DOI: 10.1007/978-3-030-24322-7_23

Title:

Optimum Number of Fourier Descriptors for Closed Boundary Retrieval

Abstract:

In the post segmentation scenario, when objects in the scene have been extracted, the focus shifts to object identification. This can be achieved through shape or texture. Finding the object boundary has been a reliable means of shape description. Among the mathematical approximation techniques for shape analysis, Fourier descriptors have proven to approximate closed boundaries of objects quite well, albeit with some limitations. A statistical thresholding technique to restrict the number of descriptors for a reasonably good approximation of the target shape is explored and tested on some medical images. Encouraging results were obtained particularly when segmentation in the preprocessing stage was effectively carried out.

Authors:

Publication:

Gate to Computer Vision and Pattern Recognition vol. 2, no. 1, pp. 1-13, 2020
DOI:10.15579/gtcvpr.0201.001013

Title:

Advanced Healthcare System using Artificial Intelligence

Abstract:

Matters like prescription and maintenance of health records hugely impact the efficiency of health care. Advanced technologies could be used to change the current scenario. This paper proposes a resourceful, web interface which enables access of medical records to the patients and a neural network model which predicts medication for ailments. The web interface acts as a medium between doctor and patient, allowing them to access the required information. The proposed method uses custom trained speech to text model and applies Natural Language Processing (NLP) on the acquired text, to provide the patient with a prescription. The proposed method further develops a neural network model which predicts the medication to be used by the patient, based on the symptoms. The proposed system achieved a Word Error Rate (WER) of 21.5% for the custom trained Speech to Text (STT) model. The AI bot used for medication prediction has achieved an accuracy of 88%.

Authors:

Publication:

2021, 11th International Conference on Cloud Computing, Data Science & Engineering (Confluence)
DOI:10.1109/Confluence51648.2021.9377084

Title:

On Efficient Shape Retrieval Of Systematic Curves Using Fourier Descriptors

Abstract:

Summarizing complex boundary shapes can be challenging. Approximation tools that already exist may require tweaking and refinement or novel approximation techniques need to be designed and tested for robustness. The problems arising in contexts like biological, satellite and medical imagery pose several challenging problems where object recognition becomes the focus of study. One of the means of recognizing objects is to describe their shapes. Boundary description is one of the means of summarizing shapes. Adding texture and color analysis to boundary is expected to take us closer to object recognition and understanding. Boundary of a shape can be modelled by spatial methods or through spectral approaches. Bilateral and radial symmetries are the popular areas of study. Since symmetry holds high level structural information, a study of symmetries can aid high level processing goals like segmentation and template matching to mention a few. The effectiveness of a near exact or exact reproduction of shape can be used for studying a variety of biological situations including leaf shape analysis and study of human parts. We focus on some shapes having bilateral symmetry, synthetic in nature, and try to retrieve a fair approximation using Fourier descriptors. We hope to extend this approach to some real-world problems where symmetries are observed. Three well known shape signatures viz. the centroid distance, area and cumulative angular functions are considered in this study. The performance of the cumulative angular function has been observed to be superior. This was tested on some shapes having bilateral symmetry and test results justify our findings.

Authors:

Publication:

International Journal on Emerging Technologies 11(3): 626-633(2020) May 2020

Title:

A two way communication system with Morse code Medium for people with Multiple Disabilities

Abstract:

This paper proposes an innovative, Morse code based two-way communication system with four user-accessible modes. The four modes are determined to provide speech to Morse vibrations and Morse to speech conversion. The proposed system is embedded with touch, gesture, vibration, microphone, and speaker modules coupled with a microprocessing unit. An S-Morse dictionary convert Morse code pattern read from the touch sensor to voice and speech from microphone to Morse vibration. A gesture recognition model using the SVM algorithm is implemented for the selection of user-accessible modes. The system modes also send and receive text data from an android application through a cloud network. The received and sent data will be converted into vibrations of Morse code pattern and voice forms respectively. The proposed method is strengthened by validating Morse conversion, machine learning algorithm accuracy, gesture recognition accuracy, Morse time analysis for alphabets, word and gesture analysis for mode selection.

Authors:

Publication:

2020 IEEE 7th Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON)
DOI: 10.1109/UPCON50219.2020.9376479

Title :

Reference Architecture for Intelligent Enterprise Solutions

Abstract:

Data in IT systems in enterprises have been growing at phenomenal pace. This has provided opportunities to run analytics to gather intelligence on key business parameters that enable them to provide better products and services to customers. While there are several Artificial Intelligence/Machine Learning (AI/ML) and Business Intelligence (BI) tools and technologies available in marketplace to run analytics, there is a need for an integrated view when developing intelligent solutions in enterprises. This paper progressively elaborates a reference model for enterprise solutions, builds an integrated view of data, information and intelligence components and presents a reference architecture for intelligent enterprise solutions. Finally, it applies the reference architecture to an insurance organization. The reference architecture is the outcome of experience and insights gathered from developing intelligent solutions for several organizations.

Authors:

Publication:

World Academy of Science, Engineering and Technology International Journal of Computer and Information Engineering Vol:15, No:7, 2021