International Research Journal on Advanced Engineering e ISSN: 2584-2854

Volume: 02
and Management Issue: 08 August 2024
https://goldncloudpublications.com Page No: 2579-2583
https://doi.org/10.47392/IRJAEM.2024.0374

Research Paper Summarizer Using AI

G. Santhoshi1, M Jyothi2, Kovvuri Ramya Sri3, G. Hasika4, G. Varsha5, R. Snigdha6
1,2
Assistant Professor, Department of Information Technology, G. Narayanamma Institute of Technology and
Science for Women, Hyderabad, India.
3Assistant Professor, Department of Computer Science Technology, G. Narayanamma Institute of Technology

and Science for Women, Hyderabad, India.

4,5,6
Student Department of Information Technology, G. Narayanamma Institute of Technology and Science for
Women, Hyderabad, India.
Emails: [email protected], [email protected] 2, [email protected],
[email protected], [email protected], [email protected]

Abstract
An AI-powered tool that can analyze and summarize research papers, making it easier for students to
understand complex academic articles. The amount of digital information is growing rapidly, making it hard
to handle and understand all the text available in different areas. It's really important to quickly and accurately
summarize large articles or research papers of text to find information, combine knowledge, and make
decisions. This research paper explains how we developed and tested a system that can turn long documents
into short, clear summaries. Develop algorithms for extracting key phrases and terms that capture the core
concepts and topics of the research paper. Develop features for Highlighting Keywords, read aloud option,
Plagiarism Check, Extracting Images, and focus areas. This tool plays an important role and help researchers
and high academic professors to get updated with the current technologies in their respective fields. The
Research Paper Summarizer Project utilizes advanced Natural Language Processing (NLP) to analyze and
summarize research papers effectively.
Keywords: Natural Language Processing (NLP), Highlighting Keywords, Read Aloud, Plagiarism, Images,
Research.

1. Introduction
In present digital world scenario, there are so many comprehend the large amount of complex
research papers that it can be challenging for information in various fields. [10]
students, researchers, and professors to keep up with 2. Literature Survey
the latest developments. Understanding and getting [1] It creates a summary by first organizing the
updated with such large amount of information takes document in layers and then choosing sentences step
lot of time, which is not always feasible. To address by step, considering what has already been included.
this issue, we developed an AI-powered tool It treats the task of picking sentences for the
specifically designed to analyze and summarize summary like a decision-making problem, where the
research papers efficiently. This tool not only document provides the information, and selecting
generates clear, efficient summaries but also each sentence is like taking an action. [2] This review
highlights key phrases and terms, reads the text on text summarization was conducted using a
aloud, checks for plagiarism, extracts relevant Systematic Literature Review (SLR) approach. SLR
images, and focuses on core concepts. We use is a method to find, evaluate, and interpret all
advanced Natural Language Processing (NLP) relevant research on a specific topic or set of research
techniques, our tool simplifies the complex academic questions. [3] The software uses the external tool
literature to simple and accessible summaries. This WordNet to improve the generated summary.
project aims to support academic and researchers by WordNet is a database that groups words by their
providing an effective solution to manage and meanings. The Natural Language Toolkit (NLTK)

IRJAEM 2579
 International Research Journal on Advanced Engineering e ISSN: 2584-2854
Volume: 02
and Management Issue: 08 August 2024
https://goldncloudpublications.com Page No: 2579-2583
https://doi.org/10.47392/IRJAEM.2024.0374

for Python is used to connect to WordNet through the view the images in the paper, to view the summary
program. The quality of the summarization is of the paper, and can also check for the plagiarism.
evaluated using ROUGE. [4] Sentence scoring [11] The flow of our system is first the user uploads
features are grouped into seven categories. One the research paper and then text processing will be
category, frequency-keyword heuristics, uses the done on the paper that is uploaded. Then it will
most common words in the document to identify its generate the summary of the paper and it will check
main themes. Sentences that include these frequent for the post-processing and removes the stop words
words are scored based on how often these words and again goes to the text processing stage. [12] This
appear. Another category, indicator phrases, focuses process will continue until it generates the
on words that usually appear in important or meaningful summary of the paper, Shown in Figure
informative parts of the text. [5] Extractive 1.
unsupervised summarization creates a summary
from a document without using any pre-labeled data
or classifications. There are three main methods to do
this: graph-based, latent variable, and term
frequency. These methods are easy to implement and
provide good results. They often produce better
outcomes compared to other advanced techniques.
[16]
3. Proposed System
In our proposed system we developed an AI
summarization tool where the users can upload a
research paper and get the summary of the paper [6].
This system is based on the research paper that is
uploaded by the users. [13] This system will generate
a summary by framing a meaningful sentence that are
extracted from the paper to generate the extract
summary for the research paper. This system will Figure 1 Architecture of The System
provide the images related to main content of the
research paper that are extracted from the paper 4. Implementation
along with summary to visualize the images the are 4.1.Natural Language Toolkit Module
present in the research paper. bIt also includes the Our system uses the nltk module. It includes text
plagiarism checker it will give the how much processing libraries for tasks such as tokenization,
percentage of text is included in the plagiarism. This stemming, lemmatization, part-of-speech tagging,
system has a read aloud feature where the users can and named entity recognition. Tokenization is used
use it to read the summary that is generated [7-9]. It to split the text form the paper into words or the
also underlines the keywords in the research paper to simple sentences. [14] Stemming is used to reduce
highlight the words in the summary. Keywords and the words to their root. Lemmatization will make
read aloud module enhance the user interaction with sure that there is no grammatical mistakes in
the paper. This module identifies the important summary. Stop words is used to remove the
words in the paper and highlight them. It will help unnecessary words from the paper like is, to, in etc.
the users to locate essential information. [17] This which will not effect to the meaning of the sentences.
system can create simplified and coherent summaries Named entity recognition will identify the proper
making complex papers more accessible and nouns in the text to add it into the summary. [15]
understandable to users. The system also allows the Also we incorporated a PIL for the image and
users to perform various controlling actions like to imageTK libraries, Image library is used for opening,

IRJAEM 2580
 International Research Journal on Advanced Engineering e ISSN: 2584-2854
Volume: 02
and Management Issue: 08 August 2024
https://goldncloudpublications.com Page No: 2579-2583
https://doi.org/10.47392/IRJAEM.2024.0374

manipulating, and saving the various images.

ImageTk will be used to display the generated
images on the TKinter GUI interface. PYPDF2
module it is used for extracting the text from the files
it can access a metadata [18]. The image extracting
is responsible for the extracting visual elements from
the research paper. [20] This system reads the reads
the text from the pdf and break down the sentences
into meaningful smaller sentences. Count the
frequency of each word to highlight the words in the
summary. Joining all the meaning full sentences to
generate a summary, shown in Figure 2. Figure 3 Result of Summary Generated

5.2. Images
Based on the paper uploaded the images present in
the paper could be charts, flowcharts or any graphs,
such images will be extracted and displayed in the
option name ‘Image’. [19], shown in Figure 4 &
Figure 5.

Figure 4 Image Generated from The Paper

Figure 2 Flow Chart for The Summarization

5. Result
5.1. Generated Summary
Firstly, the file should be uploaded (PDF). After
uploading a path will be displayed in the interface.
Later, after selecting the option named ‘Summarize
Paper’, the summary will be generated and displayed Figure 5 Image Generated from Paper
as shown in Figure 3.

IRJAEM 2581
 International Research Journal on Advanced Engineering e ISSN: 2584-2854
Volume: 02
and Management Issue: 08 August 2024
https://goldncloudpublications.com Page No: 2579-2583
https://doi.org/10.47392/IRJAEM.2024.0374

5.3. Plagiarism [7]. Dr.Annapurna P Patil,Shivan Dalmia,

It gives the plagiarism percent detected, Figure 6 Syed Abu Ayab Ansari,Tanay Aul,Varun
Clicking on the option ‘Plagiarism’. Bhatnagar “Automatic Text
Summarization”2014.
[8]. Jen-Yuan Yeh, Hao-Ren Ke,Wei-pangYang
“Text Summarization using a trainable
summarizer and latent semantic
analysis”2004.
[9]. Xiaoyue Liu,Jonathan J.,webster & Chunyu
kit “An Extractive Text Summarization
Based on significant Words”2009.
[10]. Galindo, J., & Tamayo, P. (2000). Credit risk
assessment using statistical and machine
learning: basic methodology and risk
modeling applications. Computational
Figure 6 Plagiarism Percent economics, 15, 107-143.
[11]. Ch Ramesh and K Ramya Sri, "Evaluation
Conclusion
of Machine Learning Models for Credit
This paper presents an AI tool which summarizes the
Scoring[J]", Test Engineering and
research papers and give effective summary. It
management, vol 82, Page number 2798-
utilizes Natural Language Processing (NLP)
2805.ISSN0193-4120,
techniques to give meaningful and clear summaries.
https://zenodo.org/record/7920846
In conclusion, our AI tool makes it much easier for
[12]. Kovvuri Ramya Sri,” Applied Regression
students and researchers to understand and keep up
Techniques through Cases Studies Using R”,
with the vast amount of research papers by quickly
Lambert Academic Publication, July 10,
summarizing them and highlighting key information.
2023, ISBN-10 : 6206738671, ISBN-13 :
This helps save time and improves learning and
978-6206738671
decision-making in various academic fields.
[13]. Kovvuri Ramya Sri, “Convolutional Neural
References Network- An Advanced approach of Neural
[1]. C. Virmani, A. Pillai and D. Juneja, Int. J. of
Network”, July 6, 2023, ISBN-10 :
Comp. Intell. Res. 13 621–630 (2017).
6206686531, ISBN-13 : 978-6206686538
[2]. H. P. Luhn, IBM Journal of Research and
[14]. Guna Santhoshi and Srinivasu Badugu
Development 2, 159–165 (1958)
“Development of Different Word Vectors
[3]. H. P. Edmundson, Journal of the Association
and Testing Using Text Classification
for Computing Machinery 16, 264–285
Algorithms for Telugu”, 19 April 2022, pp
(1969)
33–45, Springer, Singapore, ISBN 978-981-
[4]. Julian Kupiec,Jan Pedersen and Francine
16-9669-5 https://doi.org/10.1007/978-981-
Chen “A Trainable Document Summarizer
16-9669-5_3
2014
[15]. Santhoshi, G., et.al., published a Book with
[5]. Ishitva Awasthi ,Kuntal Gupte ,Prabjot
“Word Embedding and its Applications for
Singh Bhogal,Sahejpreet Singh
Telugu Language” VSRD academic
Anand,Prof.Piyush Kumar Soni “Natural
publishing a division of visual soft india
Language Processig based Text
pvt.ltd., Volume 1, Year 2023. ISBN-13:
Summarization” 2021
978-93-91462-86-4 Pages: vi+96
[6]. Alexander R: Fabbri, Richard
[16]. Santhoshi, G., et.al., published a paper
Socher,Dragomir Radev “SummEval: Re-
“Simulated Game Planning” in the
Evaluating Summarization”2019

IRJAEM 2582
 International Research Journal on Advanced Engineering e ISSN: 2584-2854
Volume: 02
and Management Issue: 08 August 2024
https://goldncloudpublications.com Page No: 2579-2583
https://doi.org/10.47392/IRJAEM.2024.0374

International Journal of Computer and

Internet Security. ISSN 0974-2247 Volume
15, Number 1 (2023), pp. 21-26
[17]. M.Jyothi, Sesha Bhargavi Velagaleti, A.
Vijaya Krishna, D.Sree Lakshmi
“Improving Performance of Clinical and
Operational Workflows in Health Tech
Domain using Artificial Intelligence”
International Journal for Research in
Applied Science & Engineering Technology
(IJRASET) ISSN: 2321-9653; Volume 11
Issue VI Jun 2023.
[18]. Kovvuri Ramya Sri, K Madhavi, “Machine
Learning Algorithm for Toxic Comments
Analysis” International Journal for Research
in Applied Science and Engineering
Technology, 2023, 2321-9653
https://www.ijrase t.com/ https://doi.org/10.
22214/ijraset.202 3.54217 UGC 146
[19]. Guna Santhoshi, M. Deepthi, “Real Time
Barcode Scanning without Eye Contact on
Smartphones” International Journal for
Research in Applied Science and
Engineering Technology 2023 2321-9653
https://www.ijrase t.com/ https://doi.org/10.
22214/ijraset.202 3.54219 UG.
[20]. M.Jyothi “THE ALL-IN-ONE IOT ROBOT
COMPANION” ISBN-10: 6206739457,
ISBN-13 : 978-6206739456, sep-2023.

IRJAEM 2583