Skin_Cancer.pptx

Automated Diagnosis and Prognosis of Skin Cancer Using Dermoscopy
Images with Transfer Learning Methods
Presented by
Mehebuba Nasrin Eity
ID: M200305522
Supervised by
Prof. Dr. Uzzal Kumar Acharjee
Chairman
Dept. of CSE, JnU
Dept. of Computer Science and Engineering
Jagannath University
Presentation on Degree of M.Sc in CSE
Contents
● Introduction
● Motivations
● Problem Statements
● Background Study
● Research Contributions
● Emerging key technologies, contributions and research gaps.
● System Model & Methodology
● Experimental Research and Analysis
● Conclusion
● References
2
● Skin cancer is growing exponentially nowadays because of sunlight limitations and
sunburn.
● Cancer cells are mutated cells that often grow more than normal; this abnormal growth
is called a tumor.
● There are two types of tumors: benign (stay in one part), and malignant (spread to other
parts of the body).
● Malignant tumors are cancerous because they spread to other parts of the body. On the
other hand, benign tumors stay in one part of the body.
● Benign tumors have distinct, smooth, regular borders, while malignant tumors grow
faster than benign tumors without border restriction. 3
Introduction
The motivations behind this work come out as a result of the research on the existing Skin
Cancer detection using digital “Dermoscopy Images with Transfer Learning”. It brings us to
the following conclusion.
● Data Limitation
● Lack of Early Detection Method
● Privacy on Public Health
● To build an automated system
● Environment and API support
● Depend on pre trained computer vision models
● Low accuracy gap
● High Training delay
● High Evaluation delay
4
Motivations
After research on published work, we realized a higher need using an automatic
system for dermoscopy image with transfer learning for skin cancer detection.
5
Problem Statement
In this section, I discuss the background study of CNN, EfficientNetV2B0, ViT
related to this research.
● Convolutional Neural Networks (CNN)
● EfficientNet V2 B0
● Vision Transformer (ViT)
6
Background Study
7
Background Study Cont…
● Convolutional Neural Networks (CNN)
Figure-1: CNN Architecture
8
Background Study Cont…
Figure-2: EfficientNet V2 B0 Architecture
● EfficientNet V2 B0
9
Background Study Cont…
Figure-3: ViT Architecture
● Vision Transformer (ViT)
Research contributions of this work includes:
● To design an efficient computer vision models
● To provide better classification of benign and malignant tumors.
● To mitigate against existing ISIC Archive Dataset uses the evaluate machine
learning model
● To preprocess dataset with TensorFlow's tf.data API
● To model a hierarchical based on CNN, EfficientNet V2 B0 and Vision
Transformer
● Finally, evaluate the performance of the model to solve the problem should
have a high accuracy score.
10
Research Contributions
11
Emerging key technologies, contributions and research gaps.
No
.
Authors Key Tech… Contributions and Research Gap
1 Araújo et al. (2022) CNN
need to compare effectiveness of various
model
2 Karri et al. (2023) dataset decentralized patient data
3 Daneshjou et al. (2022) ML AI supported decision making
4 Hu et al. (2023) metrics evaluate the precision of skin cancer prognosis
5 Jasil et al. (2021) VGG16, VGG19 Multi-Modal Integration
6 Mijwil et al. (2021) InceptionV3 image quality and quantity
7 SM et al. (2023) EfficientNet ethical caution of using AI
12
Emerging key technologies, contributions and research gaps.
No
.
Authors Key Tech… Contributions and Research Gap
9 Balaha (2023) CNN Evaluate the system with available dataset
10 Khan (2021) CNN need to enhance the performance
11 Afza (2022) ResNet50 improving skin lesion segmentation
12 Shorfuzzaman (2022) EfficientNetB0
need automated methods for detecting
melanoma
13
Sevli (2021)
CNN
need for accurate and efficient automated
classification
14 Keerthana (2023) CNN aiming to reduce inter-operator variability
standardization and curation to facilitate
System Model & Methodology
13
14
# Install necessary packages using pip
!pip install 
pandas  # Required for efficient data manipulation and analysis.
numpy  # Essential for numerical operations and working with arrays.
matplotlib  # Necessary for creating various types of plots and charts.
seaborn  # Enhances data visualization with attractive statistical graphics.
tensorflow  # Fundamental for building and training machine learning models.
tensorflow-addons  # Extends TensorFlow with additional functionalities.
tensorflow_hub  # Provides access to pre-trained machine learning models.
scikit-learn  # Comprehensive library for machine learning tasks.
scikit-plot  # Generates visualizations for model evaluation.
Vit-keras # Likely used for Vision Transformers in Keras for computer vision.
Python environment
15
About Dataset:
This dataset contains a balanced dataset of images of benign skin image and
malignant skin image. The data consists of two folders and pictures size (224x244).
Dataset taken from ISIC Archive.
Dataset
Malignant Benign Total
Train 1197 1440 2637
Test 300 360 660
3297
16
To achieve a faster training time we will splits train Image into train and validation.
Data augmentation layer will have to be constructed manually but to handle the
loading and passing of the image will use TensorFlow's tf.data API
Preprocessing
17
Model: "cnn_sequential_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d (Conv2D) (None, 222, 222, 16) 448
_________________________________________________________________
conv2d_1 (Conv2D) (None, 220, 220, 16) 2320
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 110, 110, 16) 0
_________________________________________________________________
conv2d_2 (Conv2D) (None, 108, 108, 8) 1160
_______________________________________________________________
conv2d_3 (Conv2D) (None, 106, 106, 8) 584
_______________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 53, 53, 8) 0
_______________________________________________________________
flatten (Flatten) (None, 22472) 0
_______________________________________________________________
dropout (Dropout) (None, 22472) 0
_______________________________________________________________
dense (Dense) (None, 128) 2876544
_______________________________________________________________
dense_1 (Dense) (None, 2) 258
=================================================================
Total params: 2,881,314
Trainable params: 2,881,314
Non-trainable params: 0
Model: "efficientnet_v2_sequential_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
efficientnet_v2_b0 (KerasLay (None, 1280) 5919312
_________________________________________________________________
dropout_1 (Dropout) (None, 1280) 0
_________________________________________________________________
dense_2 (Dense) (None, 128) 163968
_________________________________________________________________
dense_3 (Dense) (None, 2) 258
=================================================================
Total params: 6,083,538
Trainable params: 164,226
Non-trainable params: 5,919,312
Model: "vit_b16_sequential_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
vit-b16 (Functional) (None, 768) 85798656
_________________________________________________________________
dropout_2 (Dropout) (None, 768) 0
_________________________________________________________________
dense_4 (Dense) (None, 128) 98432
_________________________________________________________________
dense_5 (Dense) (None, 2) 258
=================================================================
Total params: 85,897,346
Trainable params: 98,690
Non-trainable params: 85,798,656
Trainable Parameter
18
Experimental Research and Analysis
19
The CNN model was capable of converging to a test loss similar to the observed
with the training and validation sets. We observe a stable convergence to a lower
loss. No sign of overfitting.
Results for Custom CNN
Figure : Evaluation matrix of the proposed system
20
Some overfitting have occurred during the training loss converges,, but the
validation loss is unstable for the first few epochs and converges to a higher loss
than that observed for the training set.
Results for Custom EfficientNet V2 B0
Figure : Evaluation matrics of the proposed system
21
Some slight overfitting might have occured during the last few training epochs for
the Vision Transformer model since the training and validation losses converges at
a stable rate, but towards the last few epochs there is a gap which forms between
the losses.
Results for Custom VIT B16
Figure : Evaluation matrics of the proposed system
22
CNN Confusion Matrix
=============================================
CNN Performance Metrics:
=============================================
accuracy_score: 0.8318
_____________________________________________
precision_score: 0.8372
_____________________________________________
recall_score: 0.8318
_____________________________________________
f1_score: 0.8322
_____________________________________________
matthews_corrcoef: 0.6679
=============================================
23
EfficientNet V2 B0 Confusion Matrix
=============================================
Performance Metrics:
=============================================
accuracy_score: 0.8833
_____________________________________________
precision_score: 0.8836
_____________________________________________
recall_score: 0.8833
_____________________________________________
f1_score: 0.8834
_____________________________________________
matthews_corrcoef: 0.7651
=============================================
24
VIT B16 Confusion Matrix
=============================================
Performance Metrics:
=============================================
accuracy_score: 0.8621
_____________________________________________
precision_score: 0.8642
_____________________________________________
recall_score: 0.8621
_____________________________________________
f1_score: 0.8612
_____________________________________________
matthews_corrcoef: 0.7228
=============================================
25
CNN Classification Report
precision recall f1-score support
malignant 0.78 0.87 0.83 300
benign 0.88 0.80 0.84 360
accuracy 0.83 660
macro avg 0.83 0.84 0.83 660
weighted avg 0.84 0.83 0.83 660
26
EfficientNet V2 Classification Report
precision recall f1-score support
malignant 0.87 0.88 0.87 300
benign 0.90 0.89 0.89 360
accuracy 0.88 660
macro avg 0.88 0.88 0.88 660
weighted avg 0.88 0.88 0.88 660
27
ViT-b16 Classification Report
precision recall f1-score support
malignant 0.89 0.80 0.84 300
benign 0.84 0.92 0.88 360
accuracy 0.86 660
macro avg 0.87 0.86 0.86 660
weighted avg 0.86 0.86 0.86 660
28
Performance DataFrame
Accuracy score Precision score Recall score F1 score
Matthews
corrcoef
Model CNN 0.831818 0.837174 0.831818 0.832156 0.667856
Model Efficientnet v2 0.883333 0.883583 0.883333 0.883408 0.765138
Model Vit b16 0.862121 0.864211 0.862121 0.861211 0.722836
29
Inference rate
Model: cnn_sequential_model
====================================================
Inference Time: 2.411501s ± 0.614918s
Inference Rate: 0.003654s/sample ± 0.000932s/sample
====================================================
Model: efficientnet_v2_sequential_model
====================================================
Inference Time: 3.244478s ± 1.279871s
Inference Rate: 0.004916s/sample ± 0.001939s/sample
====================================================
Model: vit_b16_sequential_model
====================================================
Inference Time: 9.293184s ± 2.437735s
Inference Rate: 0.014081s/sample ± 0.003694s/sample
30
Inference Rate vs. Matthews Correlation
31
The classification of Skin Cancer was explored with CNN, along with Transfer
Learning models such as EffiecientNet V2 and Vision Transformer. The use of
pretrained models were used to create models which were shown to outperform
the baseline CNN model.
We observe that EfficientNet V2 B0 is the best choice as it outperforms both the
baseline CNN and Vision Transformer models on the trade-off between inference
time and performance as it achieves the highest MCC while maintaining a low
inference rate.
However, Tumor is benign or malignant it should be noted that a model trained to
solve this problem should have a high accuracy score and MCC as misclassification
may bare a fatal outcome. If it is to be used in production as it will influence the
decisions made by health professionals who deal with skin cancer patients.
Conclusion
32
1. Araújo, Rafael Luz, Flávio HD de Araújo, and Romuere RV E. Silva. "Automatic segmentation of melanoma skin cancer
using transfer learning and fine-tuning." Multimedia Systems 28.4 (2022): 1239-1250.
2. Karri, Meghana, Chandra Sekhara Rao Annavarapu, and U. Rajendra Acharya. "Skin lesion segmentation using two-phase
cross-domain transfer learning framework." Computer Methods and Programs in Biomedicine 231 (2023): 107408.
3. Daneshjou, Roxana, et al. "Checklist for evaluation of image-based artificial intelligence reports in dermatology: CLEAR
derm consensus guidelines from the international skin imaging collaboration artificial intelligence working group." JAMA
dermatology 158.1 (2022): 90-96.
4. Hu, Mingzhe, Yuheng Li, and Xiaofeng Yang. "Skinsam: Empowering skin cancer segmentation with segment anything
model." arXiv preprint arXiv:2304.13973 (2023).
5. Jasil, SP Godlin, and V. Ulagamuthalvi. "Deep learning architecture using transfer learning for classification of skin
lesions." Journal of Ambient Intelligence and Humanized Computing (2021): 1-8.
6. Mijwil, Maad M. "Skin cancer disease images classification using deep learning solutions." Multimedia Tools and
Applications 80.17 (2021): 26255-26271.
7. SM, Jaisakthi, et al. "Classification of skin cancer from dermoscopic images using deep neural network architectures."
Multimedia Tools and Applications 82.10 (2023): 15763-15778.
References
33
8. Jobson, Dale, Victoria Mar, and Ian Freckelton. "Legal and ethical considerations of artificial intelligence in skin cancer
diagnosis." Australasian Journal of Dermatology 63.1 (2022): e1-e5.
9. Balaha, H.M., Hassan, A.ES. Skin cancer diagnosis based on deep transfer learning and sparrow search algorithm. Neural
Comput & Applic 35, 815–853 (2023).
10. Khan, Muhammad Attique, et al. "Pixels to classes: intelligent learning framework for multiclass skin lesion localization
and classification." Computers & Electrical Engineering 90 (2021): 106956.
11. Afza, Farhat, et al. "A hierarchical three-step superpixels and deep learning framework for skin lesion classification."
Methods 202 (2022): 88-102.
12. Shorfuzzaman, Mohammad. "An explainable stacked ensemble of deep learning models for improved melanoma skin
cancer detection." Multimedia Systems 28.4 (2022): 1309-1323.
13. Sevli, Onur. "A deep convolutional neural network-based pigmented skin lesion classification application and experts
evaluation." Neural Computing and Applications 33.18 (2021): 12039-12050.
14. Keerthana, Duggani, et al. "Hybrid convolutional neural networks with SVM classifier for classification of skin cancer."
Biomedical Engineering Advances 5 (2023): 100069.
15. Goyal, Manu, et al. "Artificial intelligence-based image classification methods for diagnosis of skin cancer: Challenges and
References
THANK YOU
34
1 sur 34

Recommandé

Blood Cell Image Classification for Detecting Malaria using CNN par
Blood Cell Image Classification for Detecting Malaria using CNNBlood Cell Image Classification for Detecting Malaria using CNN
Blood Cell Image Classification for Detecting Malaria using CNNIRJET Journal
7 vues5 diapositives
Deep Conditional Adversarial learning for polyp Segmentation par
Deep Conditional Adversarial learning for polyp SegmentationDeep Conditional Adversarial learning for polyp Segmentation
Deep Conditional Adversarial learning for polyp Segmentationmultimediaeval
90 vues15 diapositives
IRJET- A Survey on Medical Image Interpretation for Predicting Pneumonia par
IRJET- A Survey on Medical Image Interpretation for Predicting PneumoniaIRJET- A Survey on Medical Image Interpretation for Predicting Pneumonia
IRJET- A Survey on Medical Image Interpretation for Predicting PneumoniaIRJET Journal
12 vues3 diapositives
Brain Tumor Detection and Classification Using MRI Brain Images par
Brain Tumor Detection and Classification Using MRI Brain ImagesBrain Tumor Detection and Classification Using MRI Brain Images
Brain Tumor Detection and Classification Using MRI Brain ImagesIRJET Journal
12 vues7 diapositives
Dilated Inception U-Net for Nuclei Segmentation in Multi-Organ Histology Images par
Dilated Inception U-Net for Nuclei Segmentation in Multi-Organ Histology ImagesDilated Inception U-Net for Nuclei Segmentation in Multi-Organ Histology Images
Dilated Inception U-Net for Nuclei Segmentation in Multi-Organ Histology ImagesIRJET Journal
9 vues10 diapositives
A Survey of Convolutional Neural Network Architectures for Deep Learning via ... par
A Survey of Convolutional Neural Network Architectures for Deep Learning via ...A Survey of Convolutional Neural Network Architectures for Deep Learning via ...
A Survey of Convolutional Neural Network Architectures for Deep Learning via ...ijtsrd
7 vues4 diapositives

Contenu connexe

Similaire à Skin_Cancer.pptx

Prediction of Cognitive Imperiment using Deep Learning par
Prediction of Cognitive Imperiment using Deep LearningPrediction of Cognitive Imperiment using Deep Learning
Prediction of Cognitive Imperiment using Deep LearningIRJET Journal
6 vues5 diapositives
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI par
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PIDETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PIIRJET Journal
9 vues4 diapositives
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI par
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PIDETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PIIRJET Journal
7 vues4 diapositives
Performance Comparison Analysis for Medical Images Using Deep Learning Approa... par
Performance Comparison Analysis for Medical Images Using Deep Learning Approa...Performance Comparison Analysis for Medical Images Using Deep Learning Approa...
Performance Comparison Analysis for Medical Images Using Deep Learning Approa...IRJET Journal
3 vues6 diapositives
AUTOMATED WASTE MANAGEMENT SYSTEM par
AUTOMATED WASTE MANAGEMENT SYSTEMAUTOMATED WASTE MANAGEMENT SYSTEM
AUTOMATED WASTE MANAGEMENT SYSTEMIRJET Journal
5 vues5 diapositives
Skin Disease Detection using Convolutional Neural Network par
Skin Disease Detection using Convolutional Neural NetworkSkin Disease Detection using Convolutional Neural Network
Skin Disease Detection using Convolutional Neural NetworkIRJET Journal
156 vues4 diapositives

Similaire à Skin_Cancer.pptx(20)

Prediction of Cognitive Imperiment using Deep Learning par IRJET Journal
Prediction of Cognitive Imperiment using Deep LearningPrediction of Cognitive Imperiment using Deep Learning
Prediction of Cognitive Imperiment using Deep Learning
IRJET Journal6 vues
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI par IRJET Journal
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PIDETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI
IRJET Journal9 vues
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI par IRJET Journal
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PIDETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI
DETECTION OF DIFFERENT TYPES OF SKIN DISEASES USING RASPBERRY PI
IRJET Journal7 vues
Performance Comparison Analysis for Medical Images Using Deep Learning Approa... par IRJET Journal
Performance Comparison Analysis for Medical Images Using Deep Learning Approa...Performance Comparison Analysis for Medical Images Using Deep Learning Approa...
Performance Comparison Analysis for Medical Images Using Deep Learning Approa...
IRJET Journal3 vues
Skin Disease Detection using Convolutional Neural Network par IRJET Journal
Skin Disease Detection using Convolutional Neural NetworkSkin Disease Detection using Convolutional Neural Network
Skin Disease Detection using Convolutional Neural Network
IRJET Journal156 vues
CNN FEATURES ARE ALSO GREAT AT UNSUPERVISED CLASSIFICATION par cscpconf
CNN FEATURES ARE ALSO GREAT AT UNSUPERVISED CLASSIFICATION CNN FEATURES ARE ALSO GREAT AT UNSUPERVISED CLASSIFICATION
CNN FEATURES ARE ALSO GREAT AT UNSUPERVISED CLASSIFICATION
cscpconf49 vues
Brain Tumor Classification using EfficientNet Models par IRJET Journal
Brain Tumor Classification using EfficientNet ModelsBrain Tumor Classification using EfficientNet Models
Brain Tumor Classification using EfficientNet Models
IRJET Journal37 vues
Improved UNet Framework with attention for Semantic Segmentation of Tumor Reg... par IRJET Journal
Improved UNet Framework with attention for Semantic Segmentation of Tumor Reg...Improved UNet Framework with attention for Semantic Segmentation of Tumor Reg...
Improved UNet Framework with attention for Semantic Segmentation of Tumor Reg...
IRJET Journal7 vues
MIP AND UNSUPERVISED CLUSTERING FOR THE DETECTION OF BRAIN TUMOUR CELLS par AM Publications
MIP AND UNSUPERVISED CLUSTERING FOR THE DETECTION OF BRAIN TUMOUR CELLSMIP AND UNSUPERVISED CLUSTERING FOR THE DETECTION OF BRAIN TUMOUR CELLS
MIP AND UNSUPERVISED CLUSTERING FOR THE DETECTION OF BRAIN TUMOUR CELLS
AM Publications309 vues
IRJET-Multiclass Classification Method Based On Deep Learning For Leaf Identi... par IRJET Journal
IRJET-Multiclass Classification Method Based On Deep Learning For Leaf Identi...IRJET-Multiclass Classification Method Based On Deep Learning For Leaf Identi...
IRJET-Multiclass Classification Method Based On Deep Learning For Leaf Identi...
IRJET Journal49 vues
Real-Time Face-Age-Gender Detection System par IRJET Journal
Real-Time Face-Age-Gender Detection SystemReal-Time Face-Age-Gender Detection System
Real-Time Face-Age-Gender Detection System
IRJET Journal18 vues
Garbage Classification Using Deep Learning Techniques par IRJET Journal
Garbage Classification Using Deep Learning TechniquesGarbage Classification Using Deep Learning Techniques
Garbage Classification Using Deep Learning Techniques
IRJET Journal280 vues
REVIEW ON OBJECT DETECTION WITH CNN par IRJET Journal
REVIEW ON OBJECT DETECTION WITH CNNREVIEW ON OBJECT DETECTION WITH CNN
REVIEW ON OBJECT DETECTION WITH CNN
IRJET Journal9 vues
Product defect detection based on convolutional autoencoder and one-class cla... par IAESIJAI
Product defect detection based on convolutional autoencoder and one-class cla...Product defect detection based on convolutional autoencoder and one-class cla...
Product defect detection based on convolutional autoencoder and one-class cla...
IAESIJAI28 vues
Comparative Study of Pre-Trained Neural Network Models in Detection of Glaucoma par IRJET Journal
Comparative Study of Pre-Trained Neural Network Models in Detection of GlaucomaComparative Study of Pre-Trained Neural Network Models in Detection of Glaucoma
Comparative Study of Pre-Trained Neural Network Models in Detection of Glaucoma
IRJET Journal3 vues
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms par IRJET Journal
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning AlgorithmsIRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
IRJET- Diagnosis of Diabetic Retinopathy using Machine Learning Algorithms
IRJET Journal37 vues

Dernier

OBJECTIVES , CHARACTERISTICS , CONCEPT AND PHILOSOPHY OF NURSING.pptx par
OBJECTIVES , CHARACTERISTICS ,  CONCEPT AND PHILOSOPHY OF NURSING.pptxOBJECTIVES , CHARACTERISTICS ,  CONCEPT AND PHILOSOPHY OF NURSING.pptx
OBJECTIVES , CHARACTERISTICS , CONCEPT AND PHILOSOPHY OF NURSING.pptxKrishna Gandhi
10 vues52 diapositives
Fluid Administration Nursing Consideration.pptx par
Fluid Administration Nursing Consideration.pptxFluid Administration Nursing Consideration.pptx
Fluid Administration Nursing Consideration.pptxRESHMA TIRUPATI
5 vues15 diapositives
SMART RADIOLOGY : AI INNOVATIONS par
SMART RADIOLOGY  : AI INNOVATIONS SMART RADIOLOGY  : AI INNOVATIONS
SMART RADIOLOGY : AI INNOVATIONS vaarunimi
41 vues25 diapositives
Parkinson’s Disease | Athulya Assisted Living par
 Parkinson’s Disease | Athulya Assisted Living  Parkinson’s Disease | Athulya Assisted Living
Parkinson’s Disease | Athulya Assisted Living Athulya Assisted Living
14 vues8 diapositives
Renal cell carcinoma- non clear cell.pptx par
Renal cell carcinoma- non clear cell.pptxRenal cell carcinoma- non clear cell.pptx
Renal cell carcinoma- non clear cell.pptxDr. Sumit KUMAR
10 vues41 diapositives
patient consuling. par
patient consuling.patient consuling.
patient consuling.pratikshagharat
32 vues14 diapositives

Dernier(20)

OBJECTIVES , CHARACTERISTICS , CONCEPT AND PHILOSOPHY OF NURSING.pptx par Krishna Gandhi
OBJECTIVES , CHARACTERISTICS ,  CONCEPT AND PHILOSOPHY OF NURSING.pptxOBJECTIVES , CHARACTERISTICS ,  CONCEPT AND PHILOSOPHY OF NURSING.pptx
OBJECTIVES , CHARACTERISTICS , CONCEPT AND PHILOSOPHY OF NURSING.pptx
Krishna Gandhi10 vues
Fluid Administration Nursing Consideration.pptx par RESHMA TIRUPATI
Fluid Administration Nursing Consideration.pptxFluid Administration Nursing Consideration.pptx
Fluid Administration Nursing Consideration.pptx
SMART RADIOLOGY : AI INNOVATIONS par vaarunimi
SMART RADIOLOGY  : AI INNOVATIONS SMART RADIOLOGY  : AI INNOVATIONS
SMART RADIOLOGY : AI INNOVATIONS
vaarunimi41 vues
Renal cell carcinoma- non clear cell.pptx par Dr. Sumit KUMAR
Renal cell carcinoma- non clear cell.pptxRenal cell carcinoma- non clear cell.pptx
Renal cell carcinoma- non clear cell.pptx
Dr. Sumit KUMAR10 vues
Paper Napkin Manufacturing Plant Project Report 2024 par AlinaEllis1
Paper Napkin Manufacturing Plant Project Report 2024Paper Napkin Manufacturing Plant Project Report 2024
Paper Napkin Manufacturing Plant Project Report 2024
AlinaEllis15 vues
Whole Egg Powder Manufacturing Plant Project Report 2024 par AlinaEllis1
Whole Egg Powder Manufacturing Plant Project Report 2024Whole Egg Powder Manufacturing Plant Project Report 2024
Whole Egg Powder Manufacturing Plant Project Report 2024
AlinaEllis111 vues
Evovitality Revolutionizing Wellness for a Better Tomorrow.pdf par muhammadtahirbhutto9
Evovitality Revolutionizing Wellness for a Better Tomorrow.pdfEvovitality Revolutionizing Wellness for a Better Tomorrow.pdf
Evovitality Revolutionizing Wellness for a Better Tomorrow.pdf
Working-across-sectors_fulldraftreport_Aug-2019.pdf par manali9054
Working-across-sectors_fulldraftreport_Aug-2019.pdfWorking-across-sectors_fulldraftreport_Aug-2019.pdf
Working-across-sectors_fulldraftreport_Aug-2019.pdf
manali905426 vues
What's Next for OPPS: A Look at the 2024 Final Rule par Health Catalyst
What's Next for OPPS: A Look at the 2024 Final RuleWhat's Next for OPPS: A Look at the 2024 Final Rule
What's Next for OPPS: A Look at the 2024 Final Rule
Health Catalyst240 vues
Augmenting Health care delivery in Generative AI era: Balancing the hope and ... par JAI NAHAR, MD MBA
Augmenting Health care delivery in Generative AI era: Balancing the hope and ...Augmenting Health care delivery in Generative AI era: Balancing the hope and ...
Augmenting Health care delivery in Generative AI era: Balancing the hope and ...
Low Vision Managment, Age Related Macular Degeneration ARMD par mahendra singh
Low Vision Managment, Age Related Macular Degeneration ARMDLow Vision Managment, Age Related Macular Degeneration ARMD
Low Vision Managment, Age Related Macular Degeneration ARMD

Skin_Cancer.pptx

  • 1. Automated Diagnosis and Prognosis of Skin Cancer Using Dermoscopy Images with Transfer Learning Methods Presented by Mehebuba Nasrin Eity ID: M200305522 Supervised by Prof. Dr. Uzzal Kumar Acharjee Chairman Dept. of CSE, JnU Dept. of Computer Science and Engineering Jagannath University Presentation on Degree of M.Sc in CSE
  • 2. Contents ● Introduction ● Motivations ● Problem Statements ● Background Study ● Research Contributions ● Emerging key technologies, contributions and research gaps. ● System Model & Methodology ● Experimental Research and Analysis ● Conclusion ● References 2
  • 3. ● Skin cancer is growing exponentially nowadays because of sunlight limitations and sunburn. ● Cancer cells are mutated cells that often grow more than normal; this abnormal growth is called a tumor. ● There are two types of tumors: benign (stay in one part), and malignant (spread to other parts of the body). ● Malignant tumors are cancerous because they spread to other parts of the body. On the other hand, benign tumors stay in one part of the body. ● Benign tumors have distinct, smooth, regular borders, while malignant tumors grow faster than benign tumors without border restriction. 3 Introduction
  • 4. The motivations behind this work come out as a result of the research on the existing Skin Cancer detection using digital “Dermoscopy Images with Transfer Learning”. It brings us to the following conclusion. ● Data Limitation ● Lack of Early Detection Method ● Privacy on Public Health ● To build an automated system ● Environment and API support ● Depend on pre trained computer vision models ● Low accuracy gap ● High Training delay ● High Evaluation delay 4 Motivations
  • 5. After research on published work, we realized a higher need using an automatic system for dermoscopy image with transfer learning for skin cancer detection. 5 Problem Statement
  • 6. In this section, I discuss the background study of CNN, EfficientNetV2B0, ViT related to this research. ● Convolutional Neural Networks (CNN) ● EfficientNet V2 B0 ● Vision Transformer (ViT) 6 Background Study
  • 7. 7 Background Study Cont… ● Convolutional Neural Networks (CNN) Figure-1: CNN Architecture
  • 8. 8 Background Study Cont… Figure-2: EfficientNet V2 B0 Architecture ● EfficientNet V2 B0
  • 9. 9 Background Study Cont… Figure-3: ViT Architecture ● Vision Transformer (ViT)
  • 10. Research contributions of this work includes: ● To design an efficient computer vision models ● To provide better classification of benign and malignant tumors. ● To mitigate against existing ISIC Archive Dataset uses the evaluate machine learning model ● To preprocess dataset with TensorFlow's tf.data API ● To model a hierarchical based on CNN, EfficientNet V2 B0 and Vision Transformer ● Finally, evaluate the performance of the model to solve the problem should have a high accuracy score. 10 Research Contributions
  • 11. 11 Emerging key technologies, contributions and research gaps. No . Authors Key Tech… Contributions and Research Gap 1 Araújo et al. (2022) CNN need to compare effectiveness of various model 2 Karri et al. (2023) dataset decentralized patient data 3 Daneshjou et al. (2022) ML AI supported decision making 4 Hu et al. (2023) metrics evaluate the precision of skin cancer prognosis 5 Jasil et al. (2021) VGG16, VGG19 Multi-Modal Integration 6 Mijwil et al. (2021) InceptionV3 image quality and quantity 7 SM et al. (2023) EfficientNet ethical caution of using AI
  • 12. 12 Emerging key technologies, contributions and research gaps. No . Authors Key Tech… Contributions and Research Gap 9 Balaha (2023) CNN Evaluate the system with available dataset 10 Khan (2021) CNN need to enhance the performance 11 Afza (2022) ResNet50 improving skin lesion segmentation 12 Shorfuzzaman (2022) EfficientNetB0 need automated methods for detecting melanoma 13 Sevli (2021) CNN need for accurate and efficient automated classification 14 Keerthana (2023) CNN aiming to reduce inter-operator variability standardization and curation to facilitate
  • 13. System Model & Methodology 13
  • 14. 14 # Install necessary packages using pip !pip install pandas # Required for efficient data manipulation and analysis. numpy # Essential for numerical operations and working with arrays. matplotlib # Necessary for creating various types of plots and charts. seaborn # Enhances data visualization with attractive statistical graphics. tensorflow # Fundamental for building and training machine learning models. tensorflow-addons # Extends TensorFlow with additional functionalities. tensorflow_hub # Provides access to pre-trained machine learning models. scikit-learn # Comprehensive library for machine learning tasks. scikit-plot # Generates visualizations for model evaluation. Vit-keras # Likely used for Vision Transformers in Keras for computer vision. Python environment
  • 15. 15 About Dataset: This dataset contains a balanced dataset of images of benign skin image and malignant skin image. The data consists of two folders and pictures size (224x244). Dataset taken from ISIC Archive. Dataset Malignant Benign Total Train 1197 1440 2637 Test 300 360 660 3297
  • 16. 16 To achieve a faster training time we will splits train Image into train and validation. Data augmentation layer will have to be constructed manually but to handle the loading and passing of the image will use TensorFlow's tf.data API Preprocessing
  • 17. 17 Model: "cnn_sequential_model" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= conv2d (Conv2D) (None, 222, 222, 16) 448 _________________________________________________________________ conv2d_1 (Conv2D) (None, 220, 220, 16) 2320 _________________________________________________________________ max_pooling2d (MaxPooling2D) (None, 110, 110, 16) 0 _________________________________________________________________ conv2d_2 (Conv2D) (None, 108, 108, 8) 1160 _______________________________________________________________ conv2d_3 (Conv2D) (None, 106, 106, 8) 584 _______________________________________________________________ max_pooling2d_1 (MaxPooling2 (None, 53, 53, 8) 0 _______________________________________________________________ flatten (Flatten) (None, 22472) 0 _______________________________________________________________ dropout (Dropout) (None, 22472) 0 _______________________________________________________________ dense (Dense) (None, 128) 2876544 _______________________________________________________________ dense_1 (Dense) (None, 2) 258 ================================================================= Total params: 2,881,314 Trainable params: 2,881,314 Non-trainable params: 0 Model: "efficientnet_v2_sequential_model" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= efficientnet_v2_b0 (KerasLay (None, 1280) 5919312 _________________________________________________________________ dropout_1 (Dropout) (None, 1280) 0 _________________________________________________________________ dense_2 (Dense) (None, 128) 163968 _________________________________________________________________ dense_3 (Dense) (None, 2) 258 ================================================================= Total params: 6,083,538 Trainable params: 164,226 Non-trainable params: 5,919,312 Model: "vit_b16_sequential_model" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= vit-b16 (Functional) (None, 768) 85798656 _________________________________________________________________ dropout_2 (Dropout) (None, 768) 0 _________________________________________________________________ dense_4 (Dense) (None, 128) 98432 _________________________________________________________________ dense_5 (Dense) (None, 2) 258 ================================================================= Total params: 85,897,346 Trainable params: 98,690 Non-trainable params: 85,798,656 Trainable Parameter
  • 19. 19 The CNN model was capable of converging to a test loss similar to the observed with the training and validation sets. We observe a stable convergence to a lower loss. No sign of overfitting. Results for Custom CNN Figure : Evaluation matrix of the proposed system
  • 20. 20 Some overfitting have occurred during the training loss converges,, but the validation loss is unstable for the first few epochs and converges to a higher loss than that observed for the training set. Results for Custom EfficientNet V2 B0 Figure : Evaluation matrics of the proposed system
  • 21. 21 Some slight overfitting might have occured during the last few training epochs for the Vision Transformer model since the training and validation losses converges at a stable rate, but towards the last few epochs there is a gap which forms between the losses. Results for Custom VIT B16 Figure : Evaluation matrics of the proposed system
  • 22. 22 CNN Confusion Matrix ============================================= CNN Performance Metrics: ============================================= accuracy_score: 0.8318 _____________________________________________ precision_score: 0.8372 _____________________________________________ recall_score: 0.8318 _____________________________________________ f1_score: 0.8322 _____________________________________________ matthews_corrcoef: 0.6679 =============================================
  • 23. 23 EfficientNet V2 B0 Confusion Matrix ============================================= Performance Metrics: ============================================= accuracy_score: 0.8833 _____________________________________________ precision_score: 0.8836 _____________________________________________ recall_score: 0.8833 _____________________________________________ f1_score: 0.8834 _____________________________________________ matthews_corrcoef: 0.7651 =============================================
  • 24. 24 VIT B16 Confusion Matrix ============================================= Performance Metrics: ============================================= accuracy_score: 0.8621 _____________________________________________ precision_score: 0.8642 _____________________________________________ recall_score: 0.8621 _____________________________________________ f1_score: 0.8612 _____________________________________________ matthews_corrcoef: 0.7228 =============================================
  • 25. 25 CNN Classification Report precision recall f1-score support malignant 0.78 0.87 0.83 300 benign 0.88 0.80 0.84 360 accuracy 0.83 660 macro avg 0.83 0.84 0.83 660 weighted avg 0.84 0.83 0.83 660
  • 26. 26 EfficientNet V2 Classification Report precision recall f1-score support malignant 0.87 0.88 0.87 300 benign 0.90 0.89 0.89 360 accuracy 0.88 660 macro avg 0.88 0.88 0.88 660 weighted avg 0.88 0.88 0.88 660
  • 27. 27 ViT-b16 Classification Report precision recall f1-score support malignant 0.89 0.80 0.84 300 benign 0.84 0.92 0.88 360 accuracy 0.86 660 macro avg 0.87 0.86 0.86 660 weighted avg 0.86 0.86 0.86 660
  • 28. 28 Performance DataFrame Accuracy score Precision score Recall score F1 score Matthews corrcoef Model CNN 0.831818 0.837174 0.831818 0.832156 0.667856 Model Efficientnet v2 0.883333 0.883583 0.883333 0.883408 0.765138 Model Vit b16 0.862121 0.864211 0.862121 0.861211 0.722836
  • 29. 29 Inference rate Model: cnn_sequential_model ==================================================== Inference Time: 2.411501s ± 0.614918s Inference Rate: 0.003654s/sample ± 0.000932s/sample ==================================================== Model: efficientnet_v2_sequential_model ==================================================== Inference Time: 3.244478s ± 1.279871s Inference Rate: 0.004916s/sample ± 0.001939s/sample ==================================================== Model: vit_b16_sequential_model ==================================================== Inference Time: 9.293184s ± 2.437735s Inference Rate: 0.014081s/sample ± 0.003694s/sample
  • 30. 30 Inference Rate vs. Matthews Correlation
  • 31. 31 The classification of Skin Cancer was explored with CNN, along with Transfer Learning models such as EffiecientNet V2 and Vision Transformer. The use of pretrained models were used to create models which were shown to outperform the baseline CNN model. We observe that EfficientNet V2 B0 is the best choice as it outperforms both the baseline CNN and Vision Transformer models on the trade-off between inference time and performance as it achieves the highest MCC while maintaining a low inference rate. However, Tumor is benign or malignant it should be noted that a model trained to solve this problem should have a high accuracy score and MCC as misclassification may bare a fatal outcome. If it is to be used in production as it will influence the decisions made by health professionals who deal with skin cancer patients. Conclusion
  • 32. 32 1. Araújo, Rafael Luz, Flávio HD de Araújo, and Romuere RV E. Silva. "Automatic segmentation of melanoma skin cancer using transfer learning and fine-tuning." Multimedia Systems 28.4 (2022): 1239-1250. 2. Karri, Meghana, Chandra Sekhara Rao Annavarapu, and U. Rajendra Acharya. "Skin lesion segmentation using two-phase cross-domain transfer learning framework." Computer Methods and Programs in Biomedicine 231 (2023): 107408. 3. Daneshjou, Roxana, et al. "Checklist for evaluation of image-based artificial intelligence reports in dermatology: CLEAR derm consensus guidelines from the international skin imaging collaboration artificial intelligence working group." JAMA dermatology 158.1 (2022): 90-96. 4. Hu, Mingzhe, Yuheng Li, and Xiaofeng Yang. "Skinsam: Empowering skin cancer segmentation with segment anything model." arXiv preprint arXiv:2304.13973 (2023). 5. Jasil, SP Godlin, and V. Ulagamuthalvi. "Deep learning architecture using transfer learning for classification of skin lesions." Journal of Ambient Intelligence and Humanized Computing (2021): 1-8. 6. Mijwil, Maad M. "Skin cancer disease images classification using deep learning solutions." Multimedia Tools and Applications 80.17 (2021): 26255-26271. 7. SM, Jaisakthi, et al. "Classification of skin cancer from dermoscopic images using deep neural network architectures." Multimedia Tools and Applications 82.10 (2023): 15763-15778. References
  • 33. 33 8. Jobson, Dale, Victoria Mar, and Ian Freckelton. "Legal and ethical considerations of artificial intelligence in skin cancer diagnosis." Australasian Journal of Dermatology 63.1 (2022): e1-e5. 9. Balaha, H.M., Hassan, A.ES. Skin cancer diagnosis based on deep transfer learning and sparrow search algorithm. Neural Comput & Applic 35, 815–853 (2023). 10. Khan, Muhammad Attique, et al. "Pixels to classes: intelligent learning framework for multiclass skin lesion localization and classification." Computers & Electrical Engineering 90 (2021): 106956. 11. Afza, Farhat, et al. "A hierarchical three-step superpixels and deep learning framework for skin lesion classification." Methods 202 (2022): 88-102. 12. Shorfuzzaman, Mohammad. "An explainable stacked ensemble of deep learning models for improved melanoma skin cancer detection." Multimedia Systems 28.4 (2022): 1309-1323. 13. Sevli, Onur. "A deep convolutional neural network-based pigmented skin lesion classification application and experts evaluation." Neural Computing and Applications 33.18 (2021): 12039-12050. 14. Keerthana, Duggani, et al. "Hybrid convolutional neural networks with SVM classifier for classification of skin cancer." Biomedical Engineering Advances 5 (2023): 100069. 15. Goyal, Manu, et al. "Artificial intelligence-based image classification methods for diagnosis of skin cancer: Challenges and References