Scribd
Upload
868 views19 pages

Breast Cancer Classification Using Deep Learning Final

Uploaded by

user-217361
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
868 views19 pages

Breast Cancer Classification Using Deep Learning Final

Uploaded by

user-217361
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 19

PES Institute of Technology and

Management
DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING

Mini-Project Presentation
Project title: “Breast Cancer Classification with Deep
Learning”
Presented
by:
Name USN
Mythreye S R 4PM21EC061
Sadhana K R 4PM21EC081
Vaishnavi D Sharma 4PM21EC100

Project Guide Assistant Professor Project Coordinator Mrs. Prathiba


Mrs. Prathiba Assistant Professor
Head
Of
the
Depa
rtme
nt Dr.
Om
Praka
sh
Yada
v
Profe
ssor
Contents
• Introduction
• Objectives
• Problem Statement
• Literature survey
• Methadology
• Algorithm used
• Result analysis
Conclusion
Introduction
• Breast cancer is a disease where epithelial cells in the breast grow uncontrollably. This
is one of the most common cancers among women worldwide.
• Breast Cancer Classification with Deep Learning involves training artificial neural
networks to analyze medical images, such as mammograms and histopathology
images, to identify and classify breast cancer into different types, stages, and subtypes.
Objectives

 Improve the accuracy and efficiency

 Developing convolutional neural network (CNN) models

 Create a robust and reliable diagnostic tool


Problem Statement

Develop a robust and accurate machine learning model to classify breast cancer as
malignant or benign based on patient diagnostic data. The model aims to assist medical
professionals in early detection and treatment planning.
Literature Survey
Paper 1: Breast Cancer Detection using Machine Learning Approaches

Authors: Sri Hari Nallamala, Pragnyaban Mishra, Suvarna Vani Koneru


Abstract:
This paper explores the use of adaptive ensemble voting schemes for breast cancer
detection using the Wisconsin Breast Cancer dataset. It emphasizes the application
of CNN and logistic regression algorithms for high accuracy in detecting breast
cancer.
Paper 2: Breast Cancer Prediction using Deep Learning
and Machine Learning Techniques
Authors: Monika Tiwari, Rashi Bharuka, Praditi Shah, Reena Lokare
Abstract:
This study compares various machine learning and deep learning algorithms, such
as Logistic Regression, SVM, KNN, MLP, and ANN, using the Wisconsin Breast
Cancer Dataset. It concludes that deep learning models provide better accuracy and
can predict outcomes in terms of probability, which was not possible with
traditional machine learning algorithms.
Paper 3: Breast Cancer Detection and Diagnosis Using
Machine Learning: A Survey

Authors: Riyadh M. Al-Tam, Sachin M. Narangale


Abstract:
This paper reviews various techniques used for detecting and diagnosing breast cancer,
highlighting the limitations and challenges in achieving 100% accuracy due to the diverse
structure of breast cancer and different modalities used.
Paper 4: Deep Learning Applied for Histological Diagnosis
of Breast Cancer

Authors: Yasin Yari, Thuy V. Nguyen, Hieu T. Nguyen


Abstract:
This paper investigates the application of deep learning algorithms for diagnosing breast
cancer from histopathology images, achieving high accuracy in both multiclass and binary
classification modes.
Methodology
The methodology involves several key steps:

• Data Requirements: Specifications for dataset size, image resolution, format, labeling, and
data augmentation are detailed to ensure a high-quality dataset for training and evaluating the
model.

• Model Architecture: The document outlines the use of CNNs, including variants like
ResNet and Inception, for image classification. The architecture includes input and output
shapes, activation functions, optimizers, and loss functions.

• Performance Metrics: Metrics such as accuracy, precision, recall, F1-score, AUC-ROC, and
AUC-PR are used to evaluate the model's performance.
Algorithms used
1. CNN(convolution neural network)

2. Random forest

3. Logistic regressions

4. K-nearest Neighbour model(KNN)


Result analysis

Assessing Model Performance:

Evaluating the model's accuracy in classifying breast cancer images and identifying strengths and
weaknesses in its performance.

Comparative Analysis:

The model exhibits strong recall but lower precision for certain classes, indicating a tendency to
identify most true positives while also producing more false positives.
Fig 1
Have a directory for each patient ID. And in each such directory, we have the 0 and 1 directories for images
with benign and malignant content.
Fig 2
The image shows the training progress of a neural network model using TensorFlow, displaying loss and
accuracy metrics for both training and validation sets across multiple epochs.
Fig 3
The image shows the final epochs of training and the evaluation metrics of a neural network model for breast
cancer classification, including precision, recall, F1-score, accuracy, specificity, and sensitivity.
Fig 4

The graph shows the training and validation loss and accuracy for a model trained on the IDC dataset over
40 epochs, highlighting significant variability in both loss and accuracy metrics.
Conclusion

The project successfully developed a deep learning


model for the classification of breast cancer,
achieving exceptional accuracy and robustness. By
leveraging advanced deep learning techniques and a
high-quality dataset, the model demonstrated
significant potential in enhancing early detection
and improving clinical decision-making.
THANK YOU

You might also like