2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI)

Deep Learning in Mental Health: An In-depth

Analysis of Prediction Systems
2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI) | 979-8-3503-6996-0/23/$31.00 ©2023 IEEE | DOI: 10.1109/ICCSAI59793.2023.10421590

Poonam Kaushik Khushboo Bansal Yogesh Kumar

Desh Bhagat University, Desh Bhagat University, Pandit Deendayal Energy University,
Mandi Gobindgarh, India Mandi Gobindgarh, India Gandhinagar, Gujarat, India
poonam.bhargav@gmail.com erkhushboo1985@gmail.com Yogesh.arora10744@gmail.com

Abstract- Millions of people worldwide continue to suffer from Types of anxiety disorders include panic, phobias, social
a variety of diseases that frequently go undetected or untreated, anxiety, separation anxiety, illness anxiety, and obsessive-
making mental health disorders a major public health concern. In compulsive disorder. According to an epidemiology survey in
addition to a full examination of the many types of mental 2018, anxiety disorders are prevalent in 6 to 16% of the world’s
disorders, symptoms, diagnostics, and investigations of mental population [3]. Depression is characterized by pleasure loss and
disorders, this review paper gives a detailed description of the uninterrupted sadness. It influences your functioning in normal
mental disorder detection utilizing deep learning-based situations, like thinking and behavioral patterns. Depression can
methodologies. The review starts out by going over the frequency, be a chronic long-term illness or even short-term; it varies from
importance, and difficulties in timely diagnosing mental health
person to person.
illnesses. The system design for employing deep learning to identify
mental disorders early on is then further explored. In addition, the Some of the symptoms of depression include feelings of
comparative analysis component of the article, which is covered in sadness, angry outbursts, sleep disturbances, tiredness, reduced
the literature review section, provides a critical evaluation of the or increased hunger, lack of concentration, anxiety, a feeling of
datasets and evaluation metrics currently in use for training and worthlessness, suicidal thoughts, and physical pains [4].
evaluating deep learning models for the identification of mental Depression can be categorized into Major Depressive Disorder,
disorders. The advancements in this field, including the Bipolar Disorder, Seasonal Affective Disorder, Persistent
identification of depression, anxiety, schizophrenia, and other
Depressive Disorder, Psychotic Depression, and Postpartum
mental health issues, are demonstrated through a thorough
Depression. The aim of the paper is to emphasize the research
assessment of recent studies and their findings. This work also
discusses future possibilities and challenges in the field of deep conducted by various researchers in the field of mental disorders
learning for the diagnosis of mental disorders, highlighting the detection using various machine learning and deep learning
importance of large-scale, diversified, and representative datasets. approaches, based on the symptoms and diagnostics of mental
disorders [5]. The article summarizes the symptoms, diagnostic
Keywords- Mental health, Machine learning, Deep Learning, methods, and detection techniques that have previously been
Anxiety, LSTM employed for the detection of mental disorders.

I. INTRODUCTION A. Types of Mental Disorders

A mental disorder is a behavioral pattern that causes clinically Mental disorders, also referred to as psychiatric or
significant disturbance in a human being and impairment in their psychological disorders constitute a diverse array of conditions
normal functioning. It disrupts the cognitive, behavioral, that exert profound effects on an individual's cognitive processes,
physical, and emotional abilities of a human being. According to emotional states, behaviors, and their overall mental well-being (
the World Health Organization, 1 in every 8 people in the world as shown in Figure 1). The complexity of these disorders is
lives with a mental disorder. There are various types of mental reflected in the wide spectrum of symptoms, severity levels, and
disorders, such as anxiety disorders, depression, bipolar disorder, unique presentations that individuals may experience [6].
eating disorders, schizophrenia, post-traumatic stress disorder
(PTSD), and neurodevelopmental disorders [1].
These disorders are associated with high rates of morbidity,
mortality, and suicide. The main symptoms of mental disorders
are generally suicidal ideation, sleep disturbance, increased
irritability, low mood, loss of pleasure, reduced energy, and
cognitive deficit. Anxiety disorders are a result of excessive fear
and worry about forthcoming events. The main symptoms of
anxiety disorders are restlessness, low concentration, a disturbed
sleep schedule, fatigue, and unexplained pains [2].
Fig. 1. Types of mental disorder [6]

Copyright © IEEE–2023 ISBN: 979-8-3503-6996-0 364

Authorized licensed use limited to: Zhejiang University. Downloaded on July 16,2024 at 08:03:19 UTC from IEEE Xplore. Restrictions apply.
 2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI)

Anxiety Disorders: These disorders are characterized by

excessive worry, fear, or apprehension. They encompass
conditions such as Generalized Anxiety Disorder (GAD), Panic
Disorder, Social Anxiety Disorder, Specific Phobias, and
obsessive-compulsive disorder (OCD).
Mood Disorders: Mood disorders primarily impact a person's
emotional state. Included in this category are Major Depressive
Disorder (commonly known as Depression), Bipolar Disorder,
and Cyclothymic Disorder, each of which entails distinctive
patterns of mood disturbances.
Schizophrenia Spectrum and Other Psychotic Disorders:
Schizophrenia stands as a complex disorder marked by
disruptions in thought processes, perception, and behavior.
Schizoaffective Disorder, another condition in this category,
combines features of mood disorders with psychotic symptoms.
Eating Disorders: These disorders center on disordered Fig. 2. Symptoms of mental illness [9]
eating behaviors and distorted body image. Prominent examples
are Anorexia Nervosa, characterized by extreme dietary There are many different diseases that fall under the category
restriction, and Bulimia Nervosa, characterized by binge eating of mental disorders, including schizophrenia, depression, and
followed by compensatory behaviors. Binge-Eating Disorder is anxiety. Each of these ailments has a unique set of symptoms.
also notable in this category. Additionally, a person's inherent traits and past experiences have
an impact on how their mental problems emerge. As a result, due
Personality Disorders: Personality disorders involve to individual differences, two people with the same condition
enduring patterns of behavior, cognition, and inner experience may present distinct symptoms or express them in different ways
that deviate markedly from societal expectations. They include [11].
Borderline Personality Disorder, Antisocial Personality Disorder,
and narcissistic personality disorder, among others [7]. There are many different types of mental diseases, and each
is distinguished by a unique combination of symptoms. These
Somatoform Disorders: Conditions in this category involve ailments fall under the category of health problems that
physical symptoms that cannot be explained by medical predominantly affect a person's mental and emotional wellbeing.
conditions. Conversion Disorder is one such disorder, where They highlight the wide range of mental diseases by being
psychological stressors manifest as physical symptoms. extremely diversified and encompassing a wide range of
Dissociative Disorders: Dissociative disorders result in conditions [12]. They come in a wide variety of unique varieties
disruptions in consciousness, identity, memory, and awareness. rather than being restricted to a small number. Each particular
Dissociative Identity Disorder (formerly known as Multiple mental condition exhibits a distinct and recognizable pattern of
Personality Disorder) is a well-known example. symptoms that are linked to it. Each condition's specific
symptoms are unique, and they are important for identifying and
Neurocognitive Disorders: These disorders reflect understanding the problem [13].
impairments in cognitive functions such as memory, attention,
and problem-solving. Alzheimer's Disease, a progressive C. Diagnosing Mental Disorders
neurodegenerative condition, is a prime example. The diagnosis of mental disorders entails the procedure of
identifying and categorizing an individual's mental health
Impulse Control Disorders: Impulse control disorders entail condition by assessing their symptoms, behavior, and pertinent
difficulties in resisting impulses or urges. Included here are information, as depicted in Figure 3 [14].
Oppositional Defiant Disorder and Conduct Disorder, which are
often diagnosed in childhood or adolescence [8].
B. Symptoms of Mental Disorder
Depending on the specific disease and the individual, there
can be a wide variety of ways that mental problems present. As
shown in Figure 2, there is a broad range of mental diseases, each
of which is characterized by a unique set of symptoms [9]. These
symptoms act as warning signs or indicators that anything about
a person's mental health is out of the ordinary. They can manifest
in a variety of ways, such as ideas, feelings, deeds, or bodily
sensations. As a result, there are a lot of variances in how these
symptoms manifest, highlighting the fact that mental problems
do not follow a one-size-fits-all approach and can have different
effects on different people [10]. Fig. 3. Diagnosis illness-wise [14]

Copyright © IEEE–2023 ISBN: 979-8-3503-6996-0 365

Authorized licensed use limited to: Zhejiang University. Downloaded on July 16,2024 at 08:03:19 UTC from IEEE Xplore. Restrictions apply.
 2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI)

Usually, this procedure is conducted by mental health experts of elimination to arrive at the most accurate
with training, including psychiatrists, clinical psychologists, and diagnosis.
licensed therapists. The following outlines the essential elements
The length of time and the intensity of
of the mental disorder diagnostic process (Table 1): symptoms play crucial roles in the diagnosis of
mental disorders. Generally, conditions are
TABLE I. DIFFERENT WAYS TO DIAGNOSE MENTAL DISORDER [14-22] Duration and
diagnosed when symptoms endure for a
Severity
specific duration and have a substantial impact
Diagnostics Descriptions
on the individual's everyday existence.
This is the beginning of the diagnostic process,
which entails a thorough assessment of the
Continual observation and reevaluation hold
patient by a mental health professional. The
significant importance because mental health
gathering of information about the person's
conditions can evolve over time. Consistent
Clinical medical background, family history, and
Ongoing follow-up meetings with the mental health
Assessment current symptoms is part of this evaluation. To
Monitoring expert are pivotal in confirming the
gather pertinent data, the assessment typically
effectiveness of treatment and making
uses interviews, questionnaires, and
necessary modifications when required.
standardized evaluations.
Once a diagnosis is made, a treatment plan is
The DSM-5, which is published by the
developed. Treatment options may include
DSM-5 American Psychiatric Association, is
Treatment psychotherapy, medication, lifestyle changes,
(Diagnostic and frequently used by mental health professionals
Planning and other therapeutic interventions tailored to
Statistical as a reliable source for diagnosing mental
the specific disorder and individual.
Manual of disorders. The criteria and explanations it
Mental provides help professionals make accurate
During the assessment process, mental health
Disorders, Fifth diagnosis for a variety of mental health
practitioners may closely observe the
Edition) problems.
individual's conduct and mannerisms. This
Observation observation can yield valuable insights into
Certain countries utilize the International
their mental condition and the symptoms they
ICD-10 or ICD- Classification of Diseases, which is issued by
are experiencing.
11 the World Health Organization, as a tool for
(International mental disorder diagnosis. This system offers a
Some physical and neurological conditions can
Classification of worldwide standard for categorizing medical
mimic or contribute to mental health
Diseases) and mental health conditions. Rule Out
symptoms. It's essential to rule out any
Medical
underlying medical issues that may be causing
Psychometric assessments, including Conditions
or exacerbating the symptoms.
personality evaluations, cognitive
examinations, and mood measurements, might
Psychological
be conducted to acquire impartial information
Testing
about the individual's psychological II. LITERATURE REVIEW
performance. Machine and deep learning algorithms have been useful in
studying patterns and predicting the output of patient data at an
Family members, close friends, or other
relevant sources might provide further
individual level in order to achieve better results. These
perspective on the person's behavior and algorithms use statistical and computational methods that can
Collateral automatically learn from data and shows promising results in
symptoms. This additional information can be
Information
very helpful in producing a more thorough understanding large-scale mental health problems. Keeping these
review. facts in mind, we have presented the authors' contribution to
detecting and diagnosing various types of mental disorders using
Differential Mental health professionals may consider
Diagnosis multiple potential diagnoses and use a process various machine and deep learning techniques that had been
encountered during their research (Table 2):

TABLE II. COMPARATIVE ANALYSIS OF VARIOUS MENTAL DISEASES

Author’s Name Type of mental disorder Dataset Techniques Outcome
Accuracy = 93%
Ezzi et al. [1] Anxiety disorder EEG dataset CNN, LSTM Sensitivity = 95%
Specificity = 85%
Bipolar disorder corpus Accuracy = 0.858
Adaptive non-linear Recall = 0.883
Ceccarelli et al. [2] Bipolar disorder
Well being dataset judge classifier Precision = 0.867
F1 score = 0.870
Accuracy = 97%
Helaly et al. [3] Dementia disorder ADNI dataset CNN
Sensitivity =95%
Zhou et al. [4] Eating disorder Data of 123977 tweets CNN-LSTM Coherence rate = 77%
Gehricke et al. [5] Neurodevelopmental disorder Data of 72 participants Linear Regression Accuracy = 83%
Mozos et al. [6] Anxiety disorder EEG dataset of disorder Adaboost Accuracy = 79%

Copyright © IEEE–2023 ISBN: 979-8-3503-6996-0 366

Authorized licensed use limited to: Zhejiang University. Downloaded on July 16,2024 at 08:03:19 UTC from IEEE Xplore. Restrictions apply.
 2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI)

Data of 33 unipolar and 35 non-

Patel et al. [7] Depression disorder Decision tree Accuracy = 89%
depressed patients
Accuracy = 92.98%
Data based on Multi layer
Almubark et al. [8] Dementia disorder Specificity =92.68%
neuropsychological tests perceptron
Sensitivity =93.75%
Llamocca et al. [9]
Bipolar disorder Real dataset Random forest Accuracy = 88.16%
Aristizabal et al. [10] Anxiety disorder EEG dataset Neural Network Accuracy = 96%
Accuracy = 87%
Danso et al. [11] Dementia disorder SHARE dataset XgBoost Sensitivity = 76%
Specificity = 99%
EEG data of 203 healthy
Raab et al. [12] Eating disorder Random forest Accuracy = 81.25%
patients
Precision = 0.64
Laksshman et al. [13] Bipolar disorder Data consist of 1000 samples CNN Recall = 0.64

AUC = 80%
Data collected from Lithium
Nunes et al. [14] Depression disorder Random forest Sensitivity = 53%
treated patients
Specificity = 90%
Autism screening behavioural
Shamseddine et al.
Neurodevelopmental disorder data Logistic Regression Accuracy = 65%
[15]
Autism facial image data
Accuracy = 94.11%
eXtreme Gradient Recall = 94.16%
Mateo et al. [16] Bipolar disorder Real EEG data
Boosting Precision = 94.42%
F1 score = 94.79%

schizophrenia, or anxiety. It is equally crucial to establish the

III. RESEARCH METHODLOGY desired outcome of your detection system, whether it involves
To begin, the initial step is to define the problem (as depicted binary classification (determining the presence or absence of a
in Figure 4). This involves specifying the particular mental disorder) or multi-class classification (discerning among various
disorder(s) you aim to identify, such as depression, disorders).

Fig. 4. System design for mental disorder detection

Following the problem definition, the next phase is data ensuring the quality of the dataset. Tasks within this phase may
collection. During this stage, a comprehensive dataset is involve minimizing noise in the data, employing data
assembled, encompassing individuals both with and without the augmentation techniques, and normalizing data to a consistent
target mental disorder(s). It is imperative that this dataset format. Additionally, the dataset should be partitioned into
accurately represents the population you intend to serve. Each distinct subsets for training, validation, and testing purposes. This
data sample should be meticulously annotated, with labels separation is essential for the accurate evaluation of the model's
denoting the corresponding diagnosis or the severity of the performance. In the feature extraction phase, deep learning
disorder. The subsequent phase, data preprocessing, is crucial for models exhibit their capability to autonomously discern pertinent

Copyright © IEEE–2023 ISBN: 979-8-3503-6996-0 367

Authorized licensed use limited to: Zhejiang University. Downloaded on July 16,2024 at 08:03:19 UTC from IEEE Xplore. Restrictions apply.
 2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI)

patterns within the raw data. However, depending on the data's neural networks can impede clinical adoption. Researchers and
modality (e.g., text, audio, or images), there may be a necessity practitioners must continue to develop AI systems that provide
to preprocess the data and extract relevant features, such as insight into the decision-making process and are interpretable.
sentiment from text, acoustic characteristics from audio, or This review concludes by highlighting the revolutionary potential
distinctive attributes from images. The model selection process is of deep learning models for mental disorder detection. In the
critical in determining the appropriate deep learning architecture future, we can develop and investigate techniques to enhance the
to employ. Convolutional Neural Networks (CNNs) are well- interpretability of deep learning models used in mental disorder
suited for image data, while Recurrent Neural Networks (RNNs) detection. Personalized Mental Health Assessment will
excel in handling sequential data, such as text or time-series data. simultaneously investigate the creation of personalized mental
For scenarios involving multiple data modalities, architectures health assessment models.
like Convolutional-LSTM, which amalgamate features from
different sources, may be advantageous. Moreover, in cases with REFERENCES
limited data availability, exploring pre-trained models through [1] Al-Ezzi, Abdulhakim, et al. "Severity assessment of social anxiety disorder
techniques like transfer learning can be beneficial. Subsequently, using deep learning models on brain effective connectivity." IEEE
in the model development phase, you construct and train the deep Access 9 (2021): 86899-86913.
learning model utilizing the designated training dataset. It is [2] Ceccarelli, Francesco, and Marwa Mahmoud. "Multimodal temporal
machine learning for Bipolar Disorder and Depression
essential to fine-tune hyperparameters, encompassing variables Recognition." Pattern Analysis and Applications 25.3 (2022): 493-504.
such as learning rate, batch size, and model architecture, to attain [3] Helaly, Hadeer A., Mahmoud Badawy, and Amira Y. Haikal. "Deep
optimal performance. Lastly, the model's performance is assessed learning approach for early detection of Alzheimer’s disease." Cognitive
during the evaluation phase, employing pertinent metrics tailored computation (2021): 1-17.
to classification tasks, including but not limited to accuracy, [4] Zhou, Sicheng, et al. "Exploring eating disorder topics on Twitter: machine
precision, recall, F1-score, and ROC-AUC. This evaluation aids learning approach." JMIR Medical Informatics 8.10 (2020): e18273.
in gauging the model's efficacy in making accurate predictions. [5] Gehricke, Jean-G., et al. "The brain anatomy of attention-
deficit/hyperactivity disorder in young adults–a magnetic resonance
IV. CONCLUSION AND FUTURE SCOPE imaging study." PloS one 12.4 (2017): e0175433.
[6] Mozos, Oscar Martinez, et al. "Stress detection using wearable
The application of deep learning models to the detection of physiological and sociometric sensors." International journal of neural
mental disorders represents a promising strategy for enhancing systems 27.02 (2017): 1650041.
mental healthcare's early diagnosis and intervention. This review [7] Patel, Meenal J., et al. "Machine learning approaches for integrating
paper examines a vast array of studies and methodologies in this clinical and imaging features in late‐life depression classification and
burgeoning field, disclosing several significant trends and response prediction." International journal of geriatric psychiatry 30.10
takeaways. First and foremost, it is evident that deep learning (2015): 1056-1067.
techniques have demonstrated remarkable potential in accurately [8] Almubark, Ibrahim, et al. "A 5-min Cognitive Task With Deep Learning
Accurately Detects Early Alzheimer's Disease." Frontiers in aging
identifying various mental disorders, such as depression, anxiety, neuroscience 12 (2020): 603179.
schizophrenia, and bipolar disorder. These models utilise the [9] Arvindhan, M., Rajesh Kumar, D. (2022). Analysis of Load Balancing
power of neural networks to automatically extract intricate Detection Methods Using Hidden Markov Model for Secured Cloud
patterns and features from diverse data modalities, such as text, Computing Environment. In: Deepak, B.B.V.L., Parhi, D., Biswal, B.,
audio, and images, thereby facilitating more thorough and data- Jena, P.C. (eds) Applications of Computational Methods in Manufacturing
driven diagnostics. Among the significant findings is the multi- and Product Design. Lecture Notes in Mechanical Engineering. Springer,
Singapore. https://doi.org/10.1007/978-981-19-0296-3_53
modal approach, which has demonstrated superior diagnostic
[10] Aristizabal, Sara, et al. "The feasibility of wearable and self-report stress
performance by integrating information from various data detection measures in a semi-controlled lab environment." IEEE Access 9
sources. Integration of textual sentiment analysis with (2021): 102053-102068.
neuroimaging data or acoustic features from speech recordings, [11] Danso, Samuel O., et al. "Developing an explainable machine learning-
for instance, has resulted in more robust and comprehensive based personalised dementia risk prediction model: A transfer learning
assessments of mental health. Transfer learning, specifically pre- approach with ensemble learning algorithms." Frontiers in big Data 4
trained models, has emerged as a valuable asset when working (2021): 613047.
with limited datasets. Utilizing knowledge from pre-trained [12] Kumar, B., Sagar, K., Arvindhan, M., & Kr Tiwari, A. (2021). E-
models on large-scale tasks can substantially improve the Construction Cart: An Online Construction Material Ordering Service. In
2021 IEEE Mysore Sub Section International Conference (MysuruCon).
generalization capabilities of mental disorder detection models, 2021 IEEE Mysore Sub Section International Conference (MysuruCon).
even in data-scarce scenarios. Additionally, ethical IEEE. https://doi.org/10.1109/mysurucon52639.2021.9641622.
considerations are paramount in this discipline. As we advance in [13] Laksshman, Sundaram, et al. "DeepBipolar: Identifying genomic
the development of these diagnostic tools, it is crucial that we mutations for bipolar disorder via deep learning." Human mutation 38.9
maintain confidentiality, obtain informed consent, and address (2017): 1217-1224.
issues of bias and impartiality in data and algorithms. Experts in [14] Nunes, Abraham, et al. "Prediction of lithium response using clinical
deep learning and mental health professionals must work together data." Acta Psychiatrica Scandinavica 141.2 (2020): 131-141.
to achieve a balance between technological innovation and [15] Shamseddine, Hala, Safa Otoum, and Azzam Mourad. "A Federated
Learning Scheme for Neuro-developmental Disorders: Multi-Aspect ASD
ethical practice. While the accomplishments to date are Detection." arXiv preprint arXiv:2211.00643 (2022).
encouraging, there are still obstacles to surmount. The
[16] Subbulakshmi, B., Nirmala Devi, M., Sriram, Srimadhi, Arvindhan, M.
interpretability of deep learning models for mental disorder (2023). A Hybrid Machine Learning Model for House Price Prediction. In:
detection is an ongoing concern, as the 'black-box' nature of Deepak, B.B.V.L., Bahubalendruni, M.V.A.R., Parhi, D.R.K., Biswal,

Copyright © IEEE–2023 ISBN: 979-8-3503-6996-0 368

Authorized licensed use limited to: Zhejiang University. Downloaded on July 16,2024 at 08:03:19 UTC from IEEE Xplore. Restrictions apply.
 2023 International Conference on Communication, Security and Artificial Intelligence (ICCSAI)

B.B. (eds) Intelligent Manufacturing Systems in Industry 4.0. IPDIMS

2022. Lecture Notes in Mechanical Engineering. Springer, Singapore.
[17] Kaur, A. K. Sandhu and Y. Kumar, "A Hybrid Deep Transfer Learning
Approach for The Detection Of Vector-Borne Diseases," 2022 5th
International Conference on Contemporary Computing and Informatics
(IC3I), Uttar Pradesh, India, 2022, pp. 2189-2194, doi:
10.1109/IC3I56241.2022.10072576.
[18] K. Kaur, C. Singh and Y. Kumar, "Artificial Intelligence Techniques for
the Detections of Congenital Diseases: Challenges and Research
Perspectives," 2022 5th International Conference on Contemporary
Computing and Informatics (IC3I), Uttar Pradesh, India, 2022, pp. 888-
893, doi: 10.1109/IC3I56241.2022.10072469.
[19] S. Kaur, K. Bansal and Y. Kumar, "Artificial Intelligence approaches for
Predicting Hypertension Diseases: Open Challenges and Research Issues,"
2022 5th International Conference on Contemporary Computing and
Informatics (IC3I), Uttar Pradesh, India, 2022, pp. 338-343, doi:
10.1109/IC3I56241.2022.10072978.
[20] Raj, A., Jadon, S., Kulshrestha, H., Rai, V., Arvindhan, M., & Sinha, A.
(2022). Cloud Infrastructure Fault Monitoring and Prediction System using
LSTM based predictive maintenance. In 2022 10th International
Conference on Reliability, Infocom Technologies and Optimization
(Trends and Future Directions) (ICRITO). 2022 10th International
Conference on Reliability, Infocom Technologies and Optimization
(Trends and Future Directions) (ICRITO). IEEE..
[21] A. Koul, Y. Kumar and A. Gupta, "A Study on Bladder Cancer Detection
using AI-based Learning Techniques," 2022 2nd International Conference
on Technological Advancements in Computational Sciences (ICTACS),
Tashkent, Uzbekistan, 2022, pp. 600-604, doi:
10.1109/ICTACS56270.2022.9988462.
P. Bhardwaj, S. Kumar and Y. Kumar, "Deep Learning Techniques in
Gastric Cancer Prediction and Diagnosis," 2022 International Conference
on Machine Learning, Big Data, Cloud and Parallel Computing (COM-IT-
CON), Faridabad, India, 2022, pp. 843-850, doi: 10.1109/COM-IT-
CON54601.2022.9850700.

Copyright © IEEE–2023 ISBN: 979-8-3503-6996-0 369

Authorized licensed use limited to: Zhejiang University. Downloaded on July 16,2024 at 08:03:19 UTC from IEEE Xplore. Restrictions apply.