Advancing Polycystic Ovary Syndrome Detection with Artificial

Intelligence Techniques

Abir Gorrab

, Nourhène Ben Rabah

, Isuri Kariyawasam

and Bénédicte Le Grand

RIADI Laboratory, National School of Computer Science, University of Manouba, Tunisia

Centre de Recherche en Informatique, Université Paris1 Panthéon-Sorbonne, France

Keywords: Polycystic Ovary Syndrome, Artificial Intelligence, Machine Learning, Deep Learning, Medical,

Healthcare, Systematic Literature Review.

Abstract: Polycystic Ovary Syndrome (PCOS) is a common hormonal disorder that affects women of reproductive age.

Diagnosis mainly relies on traditional methods, such as clinical evaluations or laboratory tests, which can be

expensive and time-consuming and are often accompanied by complex imaging techniques. The integration

of Artificial Intelligence (AI), namely Machine Learning (ML) and Deep Learning (DL), seems to offer

promising opportunities, allowing for the analysis of large datasets to improve PCOS detection and

management. This work conducts a systematic literature review and aims to explore how ML and DL can

optimize PCOS diagnosis by analyzing the most used data and algorithms while following a rigorous

methodology to ensure the validity of the results. It also discusses the explainability of AI methods to be used

by healthcare professionals, who are always looking for reliable results to support the best possible diagnosis

for their patients.

1 INTRODUCTION

PCOS is a common hormonal disorder affecting

women of reproductive age. It is characterized by a

range of symptoms such as irregular menstrual

cycles, hyperandrogenism (excess male hormones),

the presence of multiple ovarian cysts, significant hair

loss, and notable weight gain (Narinder et al., 2023).

PCOS is often associated with ovarian dysfunctions

that can lead to miscarriages, infertility, and even

gynecological cancers. The syndrome has also a

significant financial and psychological impact on

patients. Thus, it becomes evident that the

improvement of early detection tools for PCOS are

crucial (Dana et al., 2022).

Given these major challenges for women with

PCOS and the impact on their daily lives, Artificial

Intelligence (AI) seem to offer promising prospects

for improving the understanding, diagnosis, and

management of PCOS (NMerlin and Sangeetha,

2023). Several works have addressed this emerging

research issue (Ajil et al., 2023; Srivastav et al.,

2024). Adopting AI for PCOS diagnosis seems

promising but some challenges must be addressed for

a secure and successful application.

In this work,

we propose a Systematic Literature

Review (SLR) to answer the following research issue:

How is AI presenting new perspectives for detecting

PCOS? Our article presents several contributions: We

adopted a rigorous research methodology to conduct

our SLR regarding this issue. We analyze the selected

papers deeply to handle 3 main aspects: The features

used in PCOS detection with AI, the ML, and DL

models showing the best performances and the

consideration of explainability in existing PCOS

detection works. We also give a critical eye to

existing works, highlighting the main challenges that

can give insights into future works.

The remainder of this paper is organized as

follows: We first provide background information

about PCOS diagnosis and the importance of using AI

to detect it, and previous literature reviews in Section

2, before detailing our methodology for this SLR in

Section 3. In Section 4, we present the analysis and

the results of our SLR and then discuss the challenges

associated with this research question in Section 5.

Finally, we draw our conclusions, with insights in

future works in Section 6.

Gorrab, A., Ben Rabah, N., Kariyawasam, I. and Le Grand, B.

Advancing Polycystic Ovary Syndrome Detection with Artiﬁcial Intelligence Techniques.

DOI: 10.5220/0013249800003890

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2025) - Volume 3, pages 1023-1030

ISBN: 978-989-758-737-5; ISSN: 2184-433X

1023

2 BACKGROUND AND

PREVIOUS LITERATURE

REVIEWS

In this section, we present the background of the

subject and review the literature reviews that have

addressed this topic in previous years.

2.1 Polycystic Ovary Syndrome

PCOS is the most common endocrine disorder among

women of reproductive age, affecting 10% to 15%

(Narinder et al., 2023), and, according to INSERM

(the French National Institute of Health and Medical

Research), is the leading cause of female infertility.

PCOS can be identified through imaging as an

abnormal enlargement of the ovaries (>10 ml in

volume) or many small follicles (>20 follicles smaller

than 9 mm in diameter), as shown in Figure 1.

Figure 1: Polycystic Ovary Syndrome (PCOS)

(Narinder et al., 2023).

PCOS is characterized by the ovaries producing an

abnormal number of androgens, male sex hormones

that are normally present in small quantities in

women. This overproduction of hormones can lead to

various complications, including anovulation,

causing infertility, diabetes and irregular menstrual

cycles, along with fatigue, anxiety, or depression

(Ajil et al., 2023).

 PCOS Detection

The diagnosis of PCOS is usually made by

gynecologists or midwives, who follow a step-by-

step process. The difficulty in detecting PCOS arises

from the fact that most symptoms can easily be

misattributed to puberty or stress. This is why blood

tests and laboratory analysis must be added to this

procedure to obtain results on hormone levels in the

blood. However, according to the midwives

interviewed, INSERM, and (NMerlin and Sangeetha,

2023), these results must also be combined with

imaging, such as MRI or pelvic and/or transvaginal

ultrasounds.

 Using AI to Detect PCOS:

The use of AI for the detection and management of

PCOS represents a major challenge in the medical

field. AI can accurately count the number of cysts

present in the ovaries (with a threshold of at least 20

immature follicles, each measuring less than 9 mm in

diameter, according to INSERM). Thanks to its

ability to analyze large datasets, AI enables the faster

and more precise identification of PCOS clinical

signs, thereby reducing diagnostic delays. However,

this advancement raises significant challenges,

particularly regarding the interpretability of results by

healthcare professionals.

2.2 Previous Literature Reviews

Other works studied the existing literature on the use

of AI for PCOS detection: (Barrera et al., 2023)

compared the AI algorithms used in the literature and

expressed reservations about their results due to the

heterogeneity of the available data and the risk of bias.

In general, ethical aspects and the explainability of

the algorithms were not addressed, although a

willingness to collaborate with the medical

community during validation tests is mentioned. In

their SLR, (Suha and Islam, 2023) mainly focused on

the methodology, results, recommendations, and

technical challenges related to the diagnosis of PCOS.

Even though they mentioned explainable AI, (Suha

and Islam, 2023) did not provide examples of

application in the context of PCOS, making it difficult

to assess the relevance of the proposed methods and

their potential impact on medical practice, which

limits the practical scope of the presented research.

Moreover, (Graselin et al., 2023) focused on ML

approaches for the detection of PCOS by evaluating

the effectiveness, techniques, and results of previous

studies; highlighting their technical shortcomings.

Besides, (Ahmed et al., 2023) provided an in-

depth analysis of various ML and DL approaches for

diagnosing PCOS. However, it does not address the

transparency or reproducibility of the research.

Ethical aspects and algorithm explainability were not

addressed.

In our work, we followed a rigorous research

protocol and carefully collected and analyzed research

articles to conduct an SLR dedicated to the use of AI

for PCOS detection. What distinguishes our work

from others is that we deeply focus on each aspect of

the PCOS detection process using AI. We start by

identifying the most common data and features for the

diagnosis of PCOS. We then highlight the algorithms

ICAART 2025 - 17th International Conference on Agents and Artiﬁcial Intelligence

1024

that have shown the best performance. We investigate

the explainability of AI techniques by healthcare

practitioners in the context of PCOS. We emphasize

the implications that could govern the use of AI for

PCOS diagnosis and give insight into future work.

3 RESEARCH METHODOLOGY

An SLR is a rigorous research method that

synthesizes evidence from the literature (Petersen and

Vakkalanka, 2015; Keele, 2007). In the context of

PCOS detection, this approach provides a

comprehensive and objective view of current

diagnostic methods while gathering and synthesizing

evidence from the literature.

Research Questions:

We have structured our initial research issue into 3

Research Questions (RQ):

 RQ1: What are the most used features for

diagnosing PCOS using AI?

 RQ2: Which machine learning/deep learning

algorithms have shown the best performance?

 RQ3: Is explainability considered to increase

trust in diagnostic models?

Search Query:

The following search query was executed on the

Scopus library:

(detection OR diagnostic OR diagnosis) AND (PCOS

OR "Polycystic ovary syndrome" OR "Polycystic

Ovarian Syndrome") AND ("Machine Learning" OR

ML OR "Deep Learning" OR DL).

To focus on recent works, we selected papers

published from 2017 to 2024.

Papers Selection:

To summarize the synthesis of the scientific literature,

we use the PRISMA model (Preferred Reporting

Items for Systematic Reviews and Meta-Analyses)

(Rethlefsen and Page, 2022).

Our research query returned initially n=43 articles.

We applied inclusion and exclusion criteria.

Our Inclusion criteria are: (a) Paper addressing

PCOS through machine learning, (b) Paper published

from 2017 to 2024 (to focus on recent studies), and

(a) Papers not written in English, and (b) Papers not

peer-reviewed (such as abstracts, poster, proposal,

technical reports, thesis). We had to exclude 20

articles for reasons of inaccessibility, 2 systematic

reviews, and 1 paper not written in English. We

obtained 20 papers after this first selection. The

snowballing method allowed us to integrate two new

articles into our corpus, bringing the total number of

articles to 22.

4 ANALYSIS

This section is dedicated to SLR results, presenting

responses to our three research questions.

Table 1: Overview of datasets used in studies for

diagnosing PCOS with ML.

Dataset

name

Type of data Details Ref.

PCOS

(kottarathi

l, 2018)

clinical,

metabolic,

physical,

hormonal,

ultrasound

imaging

data,

lifestyle, and

social

factors.

541 women (177

with PCOS, 364

without).

43 features.

(Ajil et al.,

2023 ;

Denny et al.,

2019 ; Modi

and Kumar,

2024 ; Subha

et al., 2024 ;

(Khanna et

al., 2023 ;

(Batra et al.,

2023 ;

Tanwar et

al., 2023)

PCOS

detection

using

ultrasound

images

(Choudhar

i, 2020)

Ultrasound

imaging

3856 ultrasound

images of women

aged 22 to 39.

Images are

classified as

infected or non-

infected.

(Diptho et

al., 2023;

NMerlin and

Sangeetha,

2023;

Rashid et al.,

2023;

Prasher and

Nelson,

2023 ;

Hosain et

al., 2022 ;

Srivastav et

al., 2024 ;

Narinder et

al., 2023)

Polycysti

c Ovary

Ultrasoun

d Images

Dataset

(Adiwija

ya and

Astuti,

2021)

Ultrasound

imaging

54 ultrasound

images from 14

PCOS patients and

40 controls. Images

were captured

using a vertical

probe ultrasound

device.

(G¨ulhan et

al., 2023)

Ovarian

ultrasoun

d image

(Dewi et

al., 2020)

Ultrasound

imaging

Contains

ultrasound images

validated by a

gynecologist,

including both

infected and non-

infected ovaries.

(Dewi et al.,

2020)

Advancing Polycystic Ovary Syndrome Detection with Artiﬁcial Intelligence Techniques

1025

4.1 What Are the Most Used Features

for Diagnosing PCOS Using AI

(RQ1)?

To answer this RQ, we analyzed the datasets used for

training and testing ML algorithms. We then

examined the different types of data used in these

models. We studied the feature selection methods

employed and conducted an in-depth analysis to

determine the features most frequently selected and

used by AI algorithms for PCOS diagnosis.

4.1.1 Dataset Analysis

We identified four datasets that we describe in Table1.

4.1.2 Data Type

Based on Table 1, datasets ‘PCOS’ and ‘PCOS

detection using ultrasound images are the most

frequently used in the literature. The first dataset

integrates clinical, metabolic, physical, hormonal,

and ultrasound imaging data and data related to

lifestyle and social factors. In contrast, the second

dataset focuses exclusively on medical imaging.

To better in-depth analysis and understanding of

the various data involved in the diagnosis and

management of PCOS, we organized the data into

three distinct categories: clinical data, non-clinical

data, and medical imaging data.

 Clinical Data:

Related to a patient's health collected during medical

care or clinical research. It can come from various

sources and is used by healthcare professionals to

make informed decisions regarding patient care. It

can include for example blood type,

Hemoglobin level,

levels of vitamin D3, heart rate per minute,

respiratory rate per minute, cycle duration, etc.

 Non-Clinical Data:

Can be derived from different aspects of lifestyle,

such as regular exercise or frequent fast-food

consumption. It may also include social factors, such

as the duration of the patient's marriage. It also

concerns physical features such as age, weight,

height, body mass index, etc.

 Medical Imaging Data:

Represents ultrasounds used to visualize the ovaries

and identify the typical cysts associated with PCOS.

4.1.3 Most Selected Features

Feature selection is a crucial step in ensuring the

performance of a ML model. It involves identifying

the most relevant variables that can effectively

distinguish whether a patient has PCOS or not.

Various statistical methods are used to select these

relevant features such as ANOVA (Analysis of

Variance), Pearson and Spearman Correlations and

Chi-Square Test. Among the non-clinical features,

‘Hair growth’ stands out as the most important

variable, followed by ‘weight gain’, ‘skin darkening’

and ‘hair loss’. The presence of ‘fast food’

consumption is also noted as a significant factor.

These symptoms are crucial for the ML model to

effectively diagnose the disease.

In terms of clinical features, the most prominent

features include ‘left follicle size’, ‘right follicle size’,

‘normal follicles on the left’, and whether ‘the cycle

is regular or irregular’.

4.2 Which AI Algorithms Have Shown

the Best Performance (RQ2)?

To detect PCOS, several AI models were used. We

differentiate these works based on the data (features)

used. Table 2 presents an overview of AI models used

for PCOS detection according to each data type.

Data preprocessing is crucial for optimizing a

model's overall performance. It ensures data quality,

making sure that the features used in ML models are

reliable and accurate. Additionally, it preserves data

integrity by consistently normalizing and scaling the

inputs (Tanwar et al., 2023). The preprocessing of

clinical and non-clinical data involves cleaning and

transforming the data to make it usable by the model.

It ensures data quality and integrity, allowing the ML

model to function optimally and produce more

accurate and reliable results. It helps correct errors

and/or remove elements that could distort the results

(NMerlin and Sangeetha, 2023; Tanwar et al., 2023).

Researchers process as well by dimensionality

reduction, helping to overcome the challenges posed

by an excess of features while retaining as much

information as possible (Denny et al., 2019). As for

data transformation, it enables the conversion of

categorical data into a numerical format that can be

used by ML algorithms, thereby ensuring fair

treatment of all variables (Batra et al., 2023).

After analyzing all these papers, we summarize

the results of the Best Performances of AI models in

terms of accuracy in Figure 2. For clinical and Non-

clinical Data, CatBoost (Modi and Kumar, 2024)

present the best accuracy. Concerning Images, the

best accuracy is given by AMCNN (Rashid et al.,

2023), Inception V3 (Narinder et al., 2023), and

CNN/ VGG16 Model (Srivastav et al., 2024], while

the INCEPTION Model and Light GBM in (NMerlin

ICAART 2025 - 17th International Conference on Agents and Artiﬁcial Intelligence

1026

and Sangeetha, 2023) give the best performance for

mixed data.

Table 2: Overview of AI models used for PCOS detection

according to each data type.

Data

type

Used algorithms References

Clinical/

Non

clinical

KNN/ NN/ NB/ SVM/

ClassificationTree

(CT)/ LR

(Dana et al., 2022)

AdaBoost/ XGBoost/

DT/ RF/ LR

(Ajil et al., 2023)

LR/ KNN/ NB/ SVM (Denny et al.,

2019

)

CatBoost/ LR/ RF/

KNN/ NB/ SVM/

AdaBoost

(Modi and Kumar,

2024)

RF/ XGB (Subha et al.,

2024

)

SHAP/ DNN (Khanna et al.,

2023)

RF/ SVM/ LR

(

Batra et al., 2023

)

RF (Tanwar et al.,

2023)

LR/ RF/ NB /KNN (Rao et al., 2024)

AdaBoost/ NB/ DT/

KNN/ SVM/

XGBoost/ RF

(Syed et al., 2023)

Images

Inception V3 (Narinder et al.,

2023)

AMCNN (Rashid et al.,

2023

)

LSTM/ BI-LSTM (Diptho et al.,

2023

)

CNN (Prasher and

Nelson, 2023;

Srinithi and

Rekha, 2023;

Srivastav and

Krishnamoorthy,

2023;

Dewi et al., 2020)

PCONet/ Inception

[(Hosain et al.,

2022)

VGG16 Model (NMerlin and

San

eetha, 2023

)

Squeeze Net/ CNN

(G¨ulhan et al.,

2023

Mixed

data

Inception Model and

Light GBM

(NMerlin and

Sangeetha, 2023)

Figure 2: Best Model Accuracy by Data Type.

4.3 Is Explainability Considered to

Increase Trust in Diagnostic

Models (RQ3)?

Most works on ML in the PCOS context focus on

algorithm performance, considered the main criterion

for comparing different methods and a key measure

of their effectiveness. This emphasis on metrics like

‘accuracy’, ‘recall’, or other performance measures

often means that many studies neglect the

transparency of results and the explainability of

model decisions.

Certain ML algorithms are referred to as ‘white-

box’, such as LR and DT; while others are referred to

as ‘black-box’, such as Deep Neural Networks, RF,

XGBoost and Adaboost. To address the limitations of

black-box models, Explainable Artificial Intelligence

(XAI) methods have been developed (Gade and Taly,

2019). In Table 3, we detail each study's use of white-

box methods as well as the black-box algorithms

employed. We also mention the name of the XAI

method, if it exists.

Table 3 reveals that only one study has addressed

the transparency of black-box models by using XAI

techniques. This study stands out for integrating

LIME (Local Interpretable Model-agnostic

Explanations) and SHAP (SHapley Additive

exPlanations), which provide interpretability to

complex models. We can also see that several

methods rely on white-box algorithms, even though

these algorithms are not the most effective. Other

studies focus solely on performance metrics without

examining interpretability, highlighting a gap in

research regarding model transparency, particularly

in critical areas such as patient care and medical

diagnosis.

Advancing Polycystic Ovary Syndrome Detection with Artiﬁcial Intelligence Techniques

1027

Table 3: Use of white-box, black-box, and XAI methods.

White-box methods

Black-box

methods

XAI

method(s)

Ref.

Linear

Discriminant

Classifier (LDT),

LR, Classification

Tree (CT), KNN,

SVM, NB

NN _ (Dana et al.,

2022)

LR, KNN, SVM,

NB, DT

AdaBoost,

XGBoost, RF

_ (Ajil et al.,

2023)

LR, Classification

and Regression

Tree (CART),

KNN, SVM, NB

RF _ (Denny et al.,

2019)

LR, KNN, SVM,

CatBoost,

RF,

AdaBoost

_ (Modi and

Kumar,

2024)

SVM RF, XGB

_ (Subha et al.,

2024)

_ DNN SHAP

LIME

(Khanna et

al., 2023)

LR, KNN, SVM,

RF _ (Batra et al.,

2023)

NB RF _ (Tanwar et

al., 2023)

LR, KNN, SVM,

RF _ (Rao et al.,

2024)

NB, KNN, SVM,

AdaBoost,

_ (Syed et al.,

2023

_ Inception V3 _ (Narinder et

al., 2023)

_ CNN, LSTM,

BI-LSTM

_ (Diptho et al.,

2023)

AMCNN

(Attention-

based Multi-

Channel

Neural

Network)

_ (Rashid et al.,

2023)

_ CNN _ (Prasher and

Nelson,

2023; Dewi

et al., 2020;

Srinithi and

Rekha, 2023;

Srivastav and

Krishnamoor

thy, 2023)

_ PCONet

(Pose-

Conditioned

Network),

Inception V3

_ (Hosain et al.,

2022)

_ CNN/VGG16 _ (Srivastav et

al., 2024)

_ Squeeze

Net/CNN

_ (G¨ulhan et

al., 2023)

_ Inception

Model and

Light GBM

_ NMerlin

and

Sangeetha,

2023)

5 DISCUSSION

By addressing our three research questions, we

identified a set of challenges that remain to be

tackled, as well as research gaps to explore. The first

challenge concerns the datasets used in the studies.

There is an urgent need to create mixed datasets that

incorporate both clinical data, non-clinical data, and

medical images. Data diversity is crucial for

improving the robustness and generalization of

diagnostic models. Currently, many studies focus on

isolated data, which may limit the models' ability to

capture all aspects of PCOS. In the same context,

why not rely on mixed features that combine different

types of data? By integrating variables from various

sources, such as biological analyses, medical

histories, and physiological characteristics, we could

improve diagnostic accuracy. The second challenge

concerns the explainability of the models. While ML

and DL algorithms can provide impressive

performance, it is equally important to ensure that

these models are interpretable and transparent.

Healthcare practitioners must be able to understand

the reasoning behind the models' predictions to

establish trust and ensure that clinical decisions are

based on solid foundations.

The lack of integrated datasets, the absence of a

multi-feature approach, and the neglect of

explainability are remaining reseach gaps,

highlighting important opportunities for future

research. Furthermore, the integration of AI in the

detection of PCOS raises critical issues related to

ethics and the regulation of health data. There are best

practice guidelines that allow the use of sensitive

data, such as health data, within an ethical framework.

These universal rules include, among others: consent,

transparency, bias prevention, anonymization, and

peer validation of a model. The consideration of

ethical concerns represents a major research gap in

PCOS detection that must be considered in future

works.

6 CONCLUSION AND FUTURE

WORKS

In this paper, we performed an SLR, describing how

AI, can present new perspectives for PCOS detection;

highlighting three aspects: Features used for PCOS

detection with AI; AI algorithms showing the best

performance; and if explainability is considered to

increase trust in diagnostic models. The advances

enabled by AI in PCOS detection are promising.

ICAART 2025 - 17th International Conference on Agents and Artiﬁcial Intelligence

1028

However, there still are many challenges: The lack of

integrated datasets, combining clinical, non-clinical

data and images; the absence of a multi-feature

approach and the lack of focus on explainability that

underscores significant opportunities for future

research. Thus, we aim to propose an enhanced PCOS

detection system, addressing the limitations of

existing works by integrating multi feature data,

focusing on XAI methods, to provide Healthcare

practitioners with interpretable results.

REFERENCES

Adiwijaya, U.,Novia, W., Untari, Astuci, W. (2021).

Polycystic ovary ultrasound images dataset.

https://dataverse.telkomuniversity.ac.id/dataset.xhtml?

persistentId=doi:10.34820/FK2/QVCP6V.

Ahmed, S., Rahman, M.S., Jahan, I., Kaiser, M.S., Mohsen,

A.S., Ghimire, D., Kim,S.H. (2023). A review on the

detection techniques of polycyctic ovary syndrome

using machine learning. IEEE Access, 11:86522-86543.

Ajil, A., Anooja, A., Ramachandra, H.V., Meenakshi, S.,

Tousif, A. (2023). Enhancing the healthcare by an

automated detection method for pcos using supervised

machine learning algorithm. In the International

Conference on Recent Advances in Information

Technology for Sustainable Development (ICRAIS).

Barrera, F., Brown, E.D., Rojo, A., Obeso, J., Plata, H,

Lincango, E.P., Shekhar, S. (2023). Application of

machine learning and artificial intelligence in the

diagnosis and classification of polycystic ovarian

syndrome- a systematic review. Frontiers in

Endocrinology, 14:1106625.

Batra, H., Saluja, K., Gupta, S., Kaushal, R., Sharma, N.,

Singh, P. (2023). Machine Learning Techniques for

Data-Driven Computer-Aided Diagnostic Method of

Polycystic Ovary Syndrome (PCOS) resulting from

Functional Ovarian Hyperandrogenism (FOH). In the

International Conference on Computational

Intelligence and Sustainable Engineering Solutions

(CISES) (pp. 195-201). IEEE.

Choudhari, A. (2020). Pcos detection using ultrasound

images.https://www.kaggle.com/datasets/anaghachoud

hari/pcos-detection-unsing-ultrasound-images.

Dana, H., Almajali, N., Alquran, H., Mustafa, W. A., Al-

Azzawi, W., Alkhayyat, A. (2022). Detection of

polycystic ovary syndrome (PCOS) using machine

learning algorithms. In the 5th international conference

on engineering technology and its applications

(IICETA), 532-536. IEEE.

Denny, A., Raj, A., Ashok, A., Ram, C. M., George, R.

(2019). i-hope: Detection and prediction system for

polycystic ovary syndrome (pcos) using machine

learning techniques. In IEEE Region 10 Conference

(TENCON), 673-678. IEEE.

Dewi, R. M., Adiwijaya, Wisesty, U. N., Jondri. (2018).

Classification of polycystic ovary based on ultrasound

images using competitive neural network. In Journal of

Physics: Conference Series, 971, 012005. IOP

Publishing.

Diptho, R. A., Jahan, N., Istiyaq, T., Anika, F., Hossain, M.

I. (2023). PCOS Diagnosis with Confluence CNN: A

Revolution in Women's Health. In the 26th

International Conference on Computer and

Information Technology (ICCIT), pp. 1-5. IEEE.

Gade, K., Geyik, S. C., Kenthapadi, K., Mithal, V., Taly, A.

(2019). Explainable AI in industry. In Proceedings of

the 25th ACM SIGKDD international conference on

knowledge discovery & data mining, 3203-3204.

Graselin, S. O., Arunprasath, T., Rajasekaran, M. P.,

Kottaimalai, R. (2023). A Systematic Review based on

the Detection of PCOS using Machine Learning

Techniques. In the 2nd International Conference on

Automation, Computing and Renewable Systems

(ICACRS), 1855-1861. IEEE.

Gulham, P.G., Ozman, G, Alptekin, H. (2023). CNN Based

Determination of Polycystic Ovarian Syndrome using

Automatic Follicle Detection Methods. Politeknik

Dergisi, 1-1.

Hagras, H. (2018). Toward human-understandable,

explainable AI. Computer, 51(9), 28-36.

Hosain, A.S., Mehedi, M.H.K., Kabir, I.E. (2022). Pconet:

A convolutional neural network architecture to detect

polycystic ovary syndrome (pcos) from ovarian

ultrasound images. In the International Conference on

Engineering and Emerging Technologies (ICEET), 1-6.

Keele, S. (2007). Guidelines for performing systematic

literature reviews in software engineering. Group

School of Computer Science and Mathematics.

Khanna, V. V., Chadaga, K., Sampathila, N., Prabhu, S.,

Bhandage, V., Hegde, G. K. (2023). A distinctive

explainable machine learning framework for detection

of polycystic ovary syndrome. Applied System

Innovation, 6(2), 32.

Kottarathil, P. (2018). Polycystic ovary syndrome (pcos).

https://www.kaggle.com/datasets/prasoonkottarathil/p

olycystic-ovary-syndrome-pcos.

Modi, N., Kumar, Y. (2024). Detection and Classification

of Polycystic Ovary Syndrome using Machine

Learning-Based Approaches. In IEEE International

Conference on Interdisciplinary Approaches in

Technology and Management for Social Innovation

(IATMSI) (Vol. 2, pp. 1-6). IEEE.

Narinder, K., Gupta, G., Kaur, P. (2023). Transfer-Based

Deep Learning Technique for PCOS Detection Using

Ultrasound Images. In the International Conference on

Network, Multimedia and Information Technology

(NMITCON) (pp. 1-6). IEEE.

NMerlin, G., Sangeetha, S., Anitha, G. (2023). An

Experimental Analysis Based on Automated Detection

of Polycystic Ovary Syndrome on Ultrasound Image

using Deep Learning Models. In the First International

Conference on Advances in Electrical, Electronics and

Computational Intelligence (ICAEECI), 1-7.

Petersen, K., Vakkalanka, S., Kuzniarz, L. (2015).

Guidelines for conducting systematic mapping studies

Advancing Polycystic Ovary Syndrome Detection with Artiﬁcial Intelligence Techniques

1029

in software engineering: An update. Information and

software technology, 64, 1-18.

Prasher, S., Nelson, L. (2023). Follicle prediction for

polycystic ovary syndrome diagnosis from ovarian

ultrasound images using cnn. In the 10th International

Conference on Computing for Sustainable Global

Development (INDIACom), 789-793. IEEE.

Rao, D., Dayma, R. R., Pendekanti, S. K. (2024). Deep

learning model for diagnosing polycystic ovary

syndrome using a comprehensive dataset from Kerala

hospitals. International Journal of Electrical &

Computer Engineering, 14(5), 2088-8708).

Rashid, S., Karnati, M., Aggarwal, G., Dutta, M. K., Sikora,

P., Burget, R. (2023). Attention-Based Multiscale Deep

Neural Network for Diagnosis of Polycystic Ovary

Syndrome Using Ovarian Ultrasound Images. In the

15th International Congress on Ultra Modern

Telecommunications and Control Systems and

Workshops (ICUMT), 44-49. IEEE.

Rethlefsen, M. L., Page, M. J. (2022). PRISMA 2020 and

PRISMA-S: common questions on tracking records and

the flow diagram. Journal of the Medical Library

Association: JMLA, 110(2), 253.

Srinithi, V., Rekha, R. (2023). Machine learning for

diagnosis of polycystic ovarian syndrome

(PCOS/PCOD). In the International Conference on

Intelligent Systems for Communication, IoT and

Security (ICISCOIS), 19-24. IEEE.

Srivastav, A. C., Krishnamoorthy, P. (2023). Unveiling

Disparities in Polycystic Ovary Syndrome Detection: A

Complex Comparative Study Integrating Deep Neural

Networks and Cutting-edge Data Visualization

Modalities. In the International Conference on Next

Generation Electronics (NEleX), 1-6. IEEE.

Srivastav, S., Guleria, K., Sharma, S. (2024). A Transfer

Learning-Based Fine Tuned VGG16 Model for PCOS

Classification. In the 2nd International Conference on

Intelligent Data Communication Technologies and

Internet of Things (IDCIoT), pp. 1074-1079. IEEE.

Stefano, B. A. (2018). Artificial intelligence, machine

learning, deep learning, and cognitive computing: what

do these terms mean and how will they impact health

care? In the Journal of arthroplasty, 33(8), 2358-2361.

Subha, R., Nayana, B. R., Radhakrishnan, R., & Sumalatha,

P. (2024). Computational intelligence for early

detection of infertility in women. Engineering

Applications of Artificial Intelligence, 127, 107400.

Suha, S. A., Islam, M. N. (2023). A systematic review and

future research agenda on detection of polycystic ovary

syndrome (PCOS) with computer-aided

techniques. Heliyon.

Syed, M. A. B., Islam, M. E., Tasnim, T. (2023).

Investigation of Polycystic Ovary Syndrome (PCOS)

Diagnosis Using Machine Learning Approaches. In the

5th International Conference on Sustainable

Technologies for Industry 5.0 (STI), pp. 1-6. IEEE.

Tanwar, A., Jain, A., Chauhan, A. (2022). Accessible

polycystic ovarian syndrome diagnosis using machine

learning. In the 3rd international conference for

emerging technology (INCET), 1-6. IEEE.

ICAART 2025 - 17th International Conference on Agents and Artiﬁcial Intelligence

1030