










































CTMJ | traditionalmedicinejournals.com Chinese Traditional Medicine Journal | 2018 | Vol 1 |Issue 5 

12 
  

Chinese Traditional Medical Journal 
 

AI-Driven Cloud-Based Deep Learning for Predictive Healthcare 

Analytics: Enhancing Disease Diagnosis with CNNs in Medical 

Imaging 

1Vijai Anand Ramar 

IBLESOFT INC, Florida, USA 

vijaianandramar@gmail.com 

2S. Rathna 

Sri Ranganathar Institute of Engineering and Technology 

Coimbatore, India. 

rathnajack@gmail.com 

 

 

1. Introduction 

The rapid advancements in Artificial Intelligence 

(AI) and Cloud Computing have revolutionized the 

healthcare industry by enabling automated and data-

driven decision-making[1]. Predictive analytics 

powered by deep learning models has significantly 

improved disease diagnosis, patient monitoring, and 

personalized treatment[2]. Among various AI 

techniques, Convolutional Neural Networks (CNNs) 

have demonstrated remarkable accuracy in 

analyzing medical images such as MRI, CT scans, 

and X-rays[3]. When integrated with cloud-based 

computing, these AI-driven models can process 

large-scale medical data efficiently, ensuring real-

time diagnosis and remote accessibility for 

healthcare providers[4]. 

 

Despite these advancements, the healthcare sector 

faces significant challenges in accurate and timely 

disease diagnosis[5]. Traditional diagnostic methods 

rely heavily on human expertise, which can lead to 

errors, delays, and inconsistent results due to fatigue 

or variability in interpretation[6]. Additionally, 

Abstract— Brain tumors pose a significant challenge in medical diagnostics, requiring early and accurate detection for 

effective treatment. Traditional diagnostic methods rely heavily on manual assessment of MRI scans, which can be time-

consuming and prone to human error. Recent advancements in deep learning and cloud computing have enabled the 

development of automated, scalable, and high-accuracy medical imaging solutions. This study presents an AI-Driven Cloud-

Based Deep Learning Framework for predictive healthcare analytics, integrating Convolutional Neural Networks (CNNs) 

with cloud computing to enhance scalability and computational efficiency. The Brain Tumor MRI Dataset from Kaggle is 

used, and extensive data preprocessing techniques such as noise reduction, normalization, and contrast enhancement are 

applied to improve image quality. The proposed model achieves 97.2% accuracy, outperforming traditional methods like 

SVM (88.5%) and Random Forest (91.3%), with an AUC-ROC score of 98.3%, demonstrating superior classification 

capability. Cloud-based deployment significantly reduces training time to 45 minutes compared to over 3 hours on local 

systems, ensuring rapid and resource-efficient processing of large MRI datasets. The findings highlight the potential of AI-

driven, cloud-integrated deep learning for real-time, scalable, and high-accuracy medical diagnostics, offering a 

transformative approach to brain tumor detection in predictive healthcare analytics. The proposed AI-driven cloud-based 

deep learning framework achieves high accuracy in brain tumor classification while significantly reducing training time 

through scalable cloud computing. This approach enhances diagnostic efficiency, making advanced medical imaging 

solutions more accessible and effective. 

 

Keywords: Brain Tumor Detection, Deep Learning, Convolutional Neural Networks (CNNs), Cloud Computing, Medical 

Imaging 

mailto:vijaianandramar@gmail.com
mailto:rathnajack@gmail.com


CTMJ | traditionalmedicinejournals.com Chinese Traditional Medicine Journal | 2018 | Vol 1 |Issue 5 

13 
  

medical imaging data is computationally intensive 

and requires significant resources for storage and 

processing[7]. Hospitals and clinics often lack the 

infrastructure to handle such large-scale data 

efficiently, leading to delayed diagnoses and 

suboptimal patient outcomes[8]. Moreover, privacy 

concerns and data security issues further hinder the 

widespread adoption of AI-driven healthcare 

solutions. 

 

To address these challenges, this research proposes 

a Cloud-Based AI-Driven Deep Learning 

Framework for predictive healthcare analytics. The 

proposed system leverages CNNs for medical image 

analysis to enhance disease diagnosis accuracy 

while utilizing cloud computing for scalable and 

efficient processing. By integrating AI models with 

cloud infrastructure, the system enables faster, more 

reliable, and remote diagnosis while ensuring secure 

data storage and real-time accessibility. This 

approach not only improves diagnostic precision and 

efficiency but also makes AI-driven healthcare 

solutions more accessible to medical institutions 

worldwide. 

 

Research Contribution 

• Introducing an AI-Driven Cloud-Based 

Deep Learning Framework that integrates 

Convolutional Neural Networks (CNNs) 

with cloud computing for enhanced 

medical imaging analysis. 

• Implementing secure cloud storage and 

encryption techniques to ensure data 

privacy and compliance with medical data 

protection standards. 

• Optimizing computational efficiency by 

leveraging cloud-based model training, 

significantly reducing training time and 

enabling scalable medical diagnostics. 

2. Literature Survey 

The integration of Artificial Intelligence (AI) and 

Cloud Computing in healthcare has gained 

significant attention due to its potential to improve 

disease diagnosis, predictive analytics, and patient 

care. Various research studies have explored the 

application of AI-driven models in medical imaging 

and healthcare analytics, while cloud-based 

frameworks have been adopted for scalable and real-

time processing of medical data[9]. This section 

reviews existing studies in AI-driven medical 

imaging, cloud-based healthcare systems, and AI-

powered predictive analytics.AI has been 

extensively used for medical imaging analysis, 

allowing for the early detection of diseases such as 

cancer, brain tumors, cardiovascular diseases, and 

lung abnormalities[10].  In this research the 

efficiency of AI-driven image analysis techniques in 

enhancing diagnostic accuracy and reducing human 

error. These studies show that AI-based models are 

capable of identifying subtle patterns in medical 

images that may not be easily detected by human 

experts[11].  

Cloud computing has emerged as a key enabler for 

scalable and remote healthcare solutions, allowing 

for the efficient storage, processing, and analysis of 

large-scale medical data[12].In this research 

demonstrates how cloud-based platforms facilitate 

real-time medical image analysis and remote 

diagnosis, especially in regions with limited 

healthcare infrastructure[13]. Cloud-based solutions 

also enable collaborative healthcare, where multiple 

medical institutions can access and analyze patient 

data securely[14]. However, latency issues, data 

privacy concerns, and integration challenges still 

pose limitations to cloud adoption in healthcare.AI-

driven predictive analytics has been widely used to 

forecast disease progression, identify at-risk 

patients, and optimize treatment plans[15]. The 

studies  have explored AI-based models that analyze 

electronic health records (EHRs), medical histories, 

and patient biomarkers to predict the likelihood of 

diseases before symptoms become severe[16]. 

These predictive models have been successful in 

improving early disease detection and personalized 

medicine. However, the effectiveness of such 

models depends on the quality of training data, 

model interpretability, and computational 

efficiency[17]. Over the years, various techniques 

have been employed in healthcare analytics for 

disease diagnosis and predictive analysis. 

Traditional machine learning methods, such as 

Support Vector Machines (SVM), Decision Trees, 

and Random Forests, have been widely used for 

medical data classification[18]. While these models 

provide decent accuracy, they often require manual 

feature extraction, making them less efficient in 

handling large-scale and complex medical data, such 

as MRI and CT scans[19]. Additionally, their 

performance is highly dependent on data quality and 

feature selection, leading to potential inconsistencies 

across different datasets[20]. 

3. Problem statement 

Although AI and cloud computing have shown great 

promise in healthcare applications, several 

challenges remain: 

• Computational Efficiency: AI-based 

medical models require high-performance 

computing resources, making them 

difficult to deploy in real-time clinical 

settings[14]. 

• Data Security & Privacy: Cloud-based 

storage of medical data raises concerns 

about cybersecurity threats, unauthorized 

access, and compliance with healthcare 

regulations (e.g., HIPAA, GDPR) [16]. 



CTMJ | traditionalmedicinejournals.com Chinese Traditional Medicine Journal | 2018 | Vol 1 |Issue 5 

14 
  

• Scalability & Integration: Many AI models 

are trained on specific datasets, limiting 

their generalization and scalability across 

diverse populations and medical 

conditions[19]. 

 

4. Methodology for AI-Driven Cloud-

Based Deep Learning in Predictive 

Healthcare Analytics 

The proposed AI-Driven Cloud-Based Deep 

Learning Framework aims to enhance disease 

diagnosis by leveraging deep learning models for 

medical imaging analysis while utilizing cloud 

computing for scalable and accuracy. 

 

Figure 1: AI-Driven Cloud-Based Deep Learning Framework for Predictive Healthcare Analytics 

The Figure 1 illustrates an AI-Driven Cloud-Based 

Deep Learning Framework for predictive healthcare 

analytics, showcasing the integration of medical 

imaging, cloud computing, and AI for accurate 

disease diagnosis. It highlights key processes such 

as data preprocessing, deep learning model training, 

cloud storage, and security measures. 

4.1 Data Collection 

The Brain Tumor MRI Dataset from Kaggle is used 

in this research, consisting of MRI scans categorized 

into Glioma Tumor, Meningioma Tumor, and No 

Tumor (Healthy Brain Images). The dataset 

provides high-quality grayscale images in standard 

formats like PNG or JPG, suitable for deep learning 

applications. Preprocessing steps include noise 

reduction, normalization, contrast enhancement, and 

data augmentation (rotation, flipping, zooming) to 

improve model performance. These refined images 

serve as input for the AI-driven cloud-based deep 

learning framework, enabling accurate and efficient 

brain tumor detection. 

 

4.2 Data Preprocessing 

To ensure high-quality input for the AI-driven 

model, the following preprocessing techniques are 

applied: 

4.2.1 Noise Reduction: Used to smooth the image 

by removing high-frequency noise while preserving 

important features are defined in Eqn. (1) 

𝐺(𝑥, 𝑦) =
1

2𝜋𝜎2
𝑒
−
𝑥2+𝑦2

2𝜎2    

  (1) 

 

 

4.2.2 Median Filtering: Each pixel is replaced 

with the median value of the surrounding pixels, 

helping remove salt-and-pepper noise. 

4.2.3 Normalization 

Min-Max Normalization: 

Ensures pixel values are scaled within a fixed 

range, typically [0,1] are in Eqn. (2): 

𝐼′ =
𝐼−𝐼min

𝐼max−𝐼min
    

  (2) 

where 𝐼 is the original pixel intensity, and 𝐼min and 

𝐼max are the minimum and maximum intensities in 

the image. 

4.2.4 Contrast Enhancement 

Histogram Equalization: 

Improves the visibility of tumor structures by 

redistributing pixel intensity values are defined in 

Eqn. (3): 

𝐻(𝑖) =
𝐶(𝑖)

𝑀𝑁
    

  (3) 

where 𝐻(𝑖) is the new intensity, 𝐶(𝑖) is the 

cumulative distribution function (CDF) of the 

intensity values, and 𝑀𝑁 is the total number of 

pixels. 

 

4.2.5 Data Augmentation 



CTMJ | traditionalmedicinejournals.com Chinese Traditional Medicine Journal | 2018 | Vol 1 |Issue 5 

15 
  

To prevent overfitting and improve generalization, 

various transformations are applied. 

• Rotation: 𝐼′(𝑥, 𝑦) = 𝐼(𝑥cos⁡ 𝜃 −
𝑦sin⁡ 𝜃, 𝑥sin⁡ 𝜃 + 𝑦cos⁡ 𝜃) 

• Scaling: 𝐼′(𝑥, 𝑦) = 𝐼(𝑠𝑥, 𝑠𝑦) 
• Flipping: Horizontal and vertical 

mirroring to increase dataset diversity. 

4.3 Deep Learning Model for Disease Diagnosis 

4.3.1 CNN Architecture Selection 

A Convolutional Neural Network (CNN) is designed 

with convolutional layers for feature extraction, 

pooling layers for dimensionality reduction, and 

fully connected layers for classification. The 

convolution operation extracts spatial patterns are 

given in eqn. (4) 

𝑍𝑖,𝑗 = ∑  𝑚 ∑  𝑛 𝐼𝑖+𝑚,𝑗+𝑛 ⋅ 𝐾𝑚,𝑛 + 𝑏  

   (4) 

 where 𝐼 is the input image, 𝐾 is the kernel, and 𝑏 is the bias. The ReLU activation function is applied⁡using⁡Eqn.⁡(5):  

𝑓(𝑥) = max(0, 𝑥)   

  (5) 

4.3.2 Model Training 

The CNN model is trained using labeled MRI 

images, optimizing weights through 

backpropagation and the Adam optimizer are given 

in Eqn. (6): 

𝑤𝑡+1 = 𝑤𝑡 − 𝜂
𝜕𝐿

𝜕𝑤
   

  (6) 

where 𝜂 is the learning rate. The Cross-Entropy Loss 

function is used for classification is defined in Eqn. 

(7): 

𝐿 = −∑ ⁡ 𝑦log⁡(�̂�)            

   (7) 

where 𝑦 is the true label and �̂� is the predicted 

probability. Model evaluation is done using 

accuracy, precision, recall, and F1-score to ensure 

robust performance. 

 

4.4 Cloud-Based Model Training 

4.4.1 Scalable Model Training Using Cloud 

Resources: Cloud platforms such as Google Cloud 

AI, AWS SageMaker, and Microsoft Azure ML 

provide scalable computing resources for deep 

learning model training. These platforms utilize 

distributed GPUs/TPUs to accelerate training: 

𝑇cloud =
𝑇local 

𝑁
    

  (8) 

where 𝑇cloud  is the training time in the cloud, 𝑇local  

is the local training time, and 𝑁 is the number of 

distributed cloud instances. This enables faster 

convergence and efficient resource utilization. 

 

4.4.2 Secure Storage and Processing of Medical 

Data: Medical images are securely stored in 

HIPAA-compliant cloud databases to ensure data 

privacy and integrity. Encryption techniques like 

AES (Advanced Encryption Standard) protect 

patient data is given in Eqn. (9): 

𝐶 = 𝐸𝑘(𝑃)    

  (9) 

where 𝐶 is the ciphertext, 𝑃 is the plaintext data, 

and 𝐸𝑘 is the encryption function using key 𝑘. 

Additionally, federated learning can be 

implemented to train models without directly 

sharing sensitive patient data. 

 

4.5 Security and Privacy Measures 

4.5.1 Data Encryption for Secure Storage and 

Transmission: Medical images and patient data are 

encrypted using Advanced Encryption Standard 

(AES) before storage or transmission. The 

encryption process ensures confidentiality is 

defined in Eqn. (10): 

𝐶 = 𝐸𝑘(𝑃)    

  (10) 

where 𝐶 is the encrypted data (ciphertext), 𝑃 is the 

original data (plaintext), and 𝐸𝑘 is the encryption 

function with key 𝑘. Secure cloud storage uses 

TLS/SSL protocols to prevent unauthorized access 

during data transfer. 

 

4.5.2 Access Control and Authentication: Role-

based access control (RBAC) is implemented to 

restrict access to sensitive medical data. 

Authentication methods such as multi-factor 

authentication (MFA) ensure only authorized 

personnel can access the system is given in Eqn. 

(11):  

𝐴 = {𝑈 ∣ 𝑈 ∈ 𝑅, 𝑃(𝑈) =  granted }  
   (11) 

where 𝐴 is the set of authorized users, 𝑅 is the role 

set, and 𝑃(𝑈) is the permission function. This 

prevents unauthorized modifications or data leaks. 

 

5. Results & Discussion 

The proposed AI-Driven Cloud-Based Deep 

Learning Model is evaluated using the Brain Tumor 

MRI Dataset. The CNN-based model achieves high 

accuracy in tumor detection, demonstrating its 

effectiveness over traditional methods. 

Additionally, the cloud-based implementation 

enhances scalability, reduces training time, and 



CTMJ | traditionalmedicinejournals.com Chinese Traditional Medicine Journal | 2018 | Vol 1 |Issue 5 

16 
  

optimizes resource utilization, making it ideal for 

real-time medical applications. 

 

Table 1: Performance Metrics of the Proposed Cloud-CNN Healthcare Model (CCHM) 

Metric Value (%) 

Accuracy 97.2 

Precision 95.8 

Recall 96.5 

F1-Score 96.1 

AUC-ROC 98.3 

The CNN model achieves a high accuracy of 97.2%, demonstrating superior tumor classification performance. 

The AUC-ROC score of 98.3% confirms the model’s strong ability to distinguish between different tumor types 

and normal cases, as shown in Table 1. 

 

Table 2: Cloud-Based Performance Metrics 

Cloud Parameter Value 

Training Time (Cloud) 45 minutes 

Inference Time 120 ms per image 

Scalability High (Handles Large Datasets) 

Storage Utilization 5 GB (Optimized for Cloud Storage) 

Cost Efficiency Reduced due to Dynamic Resource Allocation 

 

The cloud-based implementation significantly 

reduces training time (45 minutes compared to 

several hours on local machines) while maintaining 

high accuracy. The inference time is 120 ms per 

image, enabling real-time tumor classification. The 

scalability of cloud resources allows handling of 

large MRI datasets, ensuring better performance 

without hardware limitations, as given in Table 2. 

 

Table 3: Performance Comparison with Existing Methods 

Method Accuracy (%) Precision (%) Recall (%) F1-Score (%) Training Time 

Proposed 

Cloud-Based 

CNN 

97.2 95.8 96.5 96.1 45 min 

Traditional 

CNN (Local) 

94.5 92.3 93.0 92.6 3+ hours 

SVM-Based 

Model 

88.5 85.2 86.1 85.6 2 hours 

Random Forest 

Classifier 

91.3 89.8 90.5 90.1 1.5 hours 

The Proposed Cloud-Based CNN achieves the 

highest accuracy (97.2%) with significantly reduced 

training time (45 minutes) compared to the 

Traditional CNN (94.5% accuracy, 3+ hours training 

time). Compared to SVM (88.5%) and Random 

Forest (91.3%), the CNN-based approach provides 

superior precision, recall, and F1-score, ensuring 

better tumor classification, as shown in Table 3. 

 
Figure 2: Performance Comparison of Different Models for Brain Tumor Classification 



CTMJ | traditionalmedicinejournals.com Chinese Traditional Medicine Journal | 2018 | Vol 1 |Issue 5 

17 
  

Figure 2 illustrates the performance comparison of 

different models for brain tumor classification. The 

proposed Cloud-Based CNN model outperforms 

traditional approaches[21], achieving the highest 

accuracy (97.2%) with reduced training time (45 

minutes). The cloud-based implementation 

significantly enhances scalability and efficiency 

compared to local machine learning models. 

 

5.1 Discussion 

 

The proposed AI-Driven Cloud-Based Deep 

Learning Model exhibits superior performance in 

brain tumor classification compared to traditional 

methods. With an accuracy of 97.2% and an AUC-

ROC score of 98.3%, the model effectively 

differentiates between tumor types and normal 

cases. The cloud-based implementation significantly 

optimizes training efficiency, reducing training time 

to 45 minutes while maintaining high classification 

performance. Additionally, the approach enhances 

scalability, allowing efficient processing of large 

MRI datasets. Compared to SVM and Random 

Forest models, the CNN-based method achieves 

higher precision, recall, and F1-score, demonstrating 

its effectiveness in predictive healthcare analytics. 

 

6. Conclusion 

The proposed AI-Driven Cloud-Based Deep 

Learning Framework effectively enhances brain 

tumor classification using deep learning models 

integrated with cloud computing. The Cloud-Based 

CNN model achieves 97.2% accuracy, significantly 

outperforming traditional machine learning 

approaches such as SVM (88.5%) and Random 

Forest (91.3%). Cloud-based implementation 

reduces training time to 45 minutes, ensuring 

efficient processing of large MRI datasets without 

hardware limitations. The model demonstrates 

superior performance in precision (95.8%), recall 

(96.5%), and AUC-ROC (98.3%), making it a robust 

solution for predictive healthcare analytics. Future 

work will focus on integrating multi-modal medical 

data, implementing federated learning for privacy, 

enhancing model interpretability with Explainable 

AI (XAI), deploying edge computing for real-time 

diagnosis, and strengthening cloud security 

measures. 

Reference 

[1] A. K. Kalusivalingam, A. Sharma, N. Patel, and 

V. Singh, “Enhancing Remote Patient 

Monitoring Systems with Deep Learning and 

Reinforcement Learning Algorithms,” 

International Journal of AI and ML, vol. 2, no. 

10, Art. no. 10, Nov. 2013. [Online]. Available: 

https://cognitivecomputingjournal.com/index.p

hp/IJAIML-V1/article/view/118 

[2] A. K. Kalusivalingam, A. Sharma, N. Patel, and 

V. Singh, “Enhancing Mental Health 

Diagnostics: Implementing Convolutional 

Neural Networks and Natural Language 

Processing in AI-Based Assessment Tools,” 

International Journal of AI and ML, vol. 1, no. 

2, Art. no. 2, Aug. 2012. [Online]. Available: 

https://cognitivecomputingjournal.com/index.p

hp/IJAIML-V1/article/view/127 

[3] Aravindhan, K., & Subhashini, N. (2015). 

Healthcare monitoring system for elderly 

person using smart devices. Int. J. Appl. Eng. 

Res.(IJAER), 10, 20. 

[4] K. G. Srinivasa, Anupindi ,Sriram, and A. and 

Kumar, “Analytics on medical records 

collected from a distributed system deployed in 

the Indian rural demographic,” Journal of 

Management Analytics, vol. 5, no. 1, pp. 54–72, 

Jan. 2018, doi: 

10.1080/23270012.2018.1433085. 

[5] Y. Zhang, M. Qiu, C.-W. Tsai, M. M. Hassan, 

and A. Alamri, “Health-CPS: Healthcare 

Cyber-Physical System Assisted by Cloud and 

Big Data,” IEEE Systems Journal, vol. 11, no. 

1, pp. 88–95, Mar. 2017, doi: 

10.1109/JSYST.2015.2460747. 

[6] H. Alharthi, “Healthcare predictive analytics: 

An overview with a focus on Saudi Arabia,” 

Journal of Infection and Public Health, vol. 11, 

no. 6, pp. 749–756, Nov. 2018, doi: 

10.1016/j.jiph.2018.02.005. 

[7] S. A. Choudhry, J. Li, D. Davis, C. Erdmann, 

R. Sikka, and B. Sutariya, “A Public-Private 

Partnership Develops and Externally Validates 

a 30-Day Hospital Readmission Risk 

Prediction Model,” OJPHI, vol. 5, no. 2, Jun. 

2013, doi: 10.5210/ojphi.v5i2.4726. 

[8] D. E. Jones, M. Greenberg, and M. Crowley, 

“Early Social-Emotional Functioning and 

Public Health: The Relationship Between 

Kindergarten Social Competence and Future 

Wellness,” Am J Public Health, vol. 105, no. 

11, pp. 2283–2290, Nov. 2015, doi: 

10.2105/AJPH.2015.302630. 

[9] J. Kaur and Dr. Kulwinder Singh Mann, “AI 

based HealthCare Platform for Real Time, 

Predictive and Prescriptive Analytics using 

Reactive Programming,” J. Phys.: Conf. Ser., 

vol. 933, p. 012010, Jan. 2018, doi: 

10.1088/1742-6596/933/1/012010. 

[10] K. T. Chui, W. Alhalabi, S. S. H. Pang, P. O. de 

Pablos, R. W. Liu, and M. Zhao, “Disease 

Diagnosis in Smart Healthcare: Innovation, 

Technologies and Applications,” 

Sustainability, vol. 9, no. 12, Art. no. 12, Dec. 

2017, doi: 10.3390/su9122309. 

[11] B. Nithya and V. Ilango, “Predictive analytics 

in health care using machine learning tools and 

techniques,” in 2017 International Conference 

on Intelligent Computing and Control Systems 



CTMJ | traditionalmedicinejournals.com Chinese Traditional Medicine Journal | 2018 | Vol 1 |Issue 5 

18 
  

(ICICCS), Jun. 2017, pp. 492–499. doi: 

10.1109/ICCONS.2017.8250771. 

[12] A. Sheth, U. Jaimini, K. Thirunarayan, and T. 

Banerjee, “Augmented personalized health: 

How smart data with IoTs and AI is about to 

change healthcare,” in 2017 IEEE 3rd 

International Forum on Research and 

Technologies for Society and Industry (RTSI), 

Sep. 2017, pp. 1–6. doi: 

10.1109/RTSI.2017.8065963. 

[13] P. Mamoshina et al., “Converging blockchain 

and next-generation artificial intelligence 

technologies to decentralize and accelerate 

biomedical research and healthcare,” 

Oncotarget, vol. 9, no. 5, pp. 5665–5690, Nov. 

2017, doi: 10.18632/oncotarget.22345. 

[14] S. Mahmud, R. Iqbal, and F. Doctor, “Cloud 

enabled data analytics and visualization 

framework for health-shocks prediction,” 

Future Generation Computer Systems, vol. 65, 

pp. 169–181, Dec. 2016, doi: 

10.1016/j.future.2015.10.014. 

[15] Y. Ma, Y. Wang, J. Yang, Y. Miao, and W. Li, 

“Big Health Application System based on 

Health Internet of Things and Big Data,” IEEE 

Access, vol. 5, pp. 7885–7897, 2017, doi: 

10.1109/ACCESS.2016.2638449. 

[16] N. Abirami, S. Karthik, and M. Kanimozhi, 

“AUTOMATED BRAIN TUMOR 

DETECTION AND IDENTIFICATION 

USING MEDICAL IMAGING,” 

INTERNATIONAL JOURNAL OF RESEARCH 

IN COMPUTER APPLICATIONS AND 

ROBOTICS, no. 9, 2015. 

[17] M. A. Khan et al., “Brain tumor detection and 

classification: A framework of marker‐based 

watershed algorithm and multilevel priority 

features selection,” Microscopy Res & 

Technique, vol. 82, no. 6, pp. 909–922, Jun. 

2019, doi: 10.1002/jemt.23238. 

[18] Veerashetty, S. (2014). Multi-modal weighted 

denoising coder for the management of lost 

information in healthcare big data. Education, 

2019. 

[19] Carneiro, G., Peng, T., Bayer, C., & Navab, N. 

(2017). Automatic quantification of tumour 

hypoxia from multi-modal microscopy images 

using weakly-supervised learning methods. 

IEEE transactions on medical imaging, 36(7), 

1405-1417. 

[20] D. PeterAugustine, “Enhancing the Efficiency 

of Parallel Genetic Algorithms for Medical 

Image Processing with Hadoop,” IJCA, vol. 

108, no. 17, pp. 11–16, Dec. 2014, doi: 

10.5120/19002-0483. 

[21] M. H. Ghahramani, M. Zhou, and C. T. Hon, 

“Toward cloud computing QoS architecture: 

analysis of cloud systems and cloud services,” 

IEEE/CAA Journal of Automatica Sinica, vol. 

4, no. 1, pp. 6–18, Jan. 2017, doi: 

10.1109/JAS.2017.7510313. 

 


