Scribd Logo
Upload

Welcome to Scribd!

  • 51 views

    Breast Cancer MET

    Uploaded by

    vishalmate10
    This document outlines a final year project that aims to develop a deep learning model for breast cancer classification and detection using convolutional neural networks (CNNs). The project uses the BreakHis histopathology image dataset containing benign and malignant breast cancer cases. A custom CNN architecture is designed and trained on preprocessed images from the dataset. Techniques like dropout and batch normalization are employed to improve the model's performance. The CNN learns visual features to distinguish between benign and malignant cases. The results demonstrate CNNs can accurately analyze histopathology images and aid in early breast cancer detection.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Breast Cancer MET

    Uploaded by

    vishalmate10
    51 views24 pages
    This document outlines a final year project that aims to develop a deep learning model for breast cancer classification and detection using convolutional neural networks (CNNs). The project uses the BreakHis histopathology image dataset containing benign and malignant breast cancer cases. A custom CNN architecture is designed and trained on preprocessed images from the dataset. Techniques like dropout and batch normalization are employed to improve the model's performance. The CNN learns visual features to distinguish between benign and malignant cases. The results demonstrate CNNs can accurately analyze histopathology images and aid in early breast cancer detection.

    Original Description:

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document outlines a final year project that aims to develop a deep learning model for breast cancer classification and detection using convolutional neural networks (CNNs). The project uses the BreakHis histopathology image dataset containing benign and malignant breast cancer cases. A custom CNN architecture is designed and trained on preprocessed images from the dataset. Techniques like dropout and batch normalization are employed to improve the model's performance. The CNN learns visual features to distinguish between benign and malignant cases. The results demonstrate CNNs can accurately analyze histopathology images and aid in early breast cancer detection.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    51 views24 pages

    Breast Cancer MET

    Uploaded by

    vishalmate10
    This document outlines a final year project that aims to develop a deep learning model for breast cancer classification and detection using convolutional neural networks (CNNs). The project uses the BreakHis histopathology image dataset containing benign and malignant breast cancer cases. A custom CNN architecture is designed and trained on preprocessed images from the dataset. Techniques like dropout and batch normalization are employed to improve the model's performance. The CNN learns visual features to distinguish between benign and malignant cases. The results demonstrate CNNs can accurately analyze histopathology images and aid in early breast cancer detection.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    of 24

    Bhujbal Knowledge City

    MET Institute of Engineering


    Department of Information Technology

    Final Year Project

    Breast Cancer Classification and Detection


    using CNN
    A Project by:

    Sanskruti Bagul
    Aaditi Andore
    Siddhika Ghule
    Karan Kawre

    Project Guide
    Dr. R. P. Dahake
    Content
    •Introduction
    •Literature Survey
    •Problem Statement
    •Objectives
    •Proposed System
    •Dataset Used
    •Algorithm used
    •Software Requirements
    •Result and Analysis
    •Conclusion and Future scope
    •References
    Introduction
    • Breast cancer is the country's leading cause of cancer in women, followed by
    cervical cancer. Together they account for 39.4 percent of the total cancer
    cases in women in India in 2020.
    • Despite High Awareness, 75% of Indian Women Shy Away from Breast
    Cancer Screening.
    • In 2020, more than two lakh women in India were estimated to have been
    diagnosed with breast cancer, and more than 76,000 deaths were reported as
    per the estimates.
    • As per the 2020 National Cancer Registry Program Report, the number is
    expected to rise to more than 2.3 lakh cases in 2025.
    Literature Review
    G. Chen, Y. Chen, Z. Yuan, X. Lu, X. Zhu, and W. Li, "Breast Cancer
    Image Classification based on CNN and Bit-Plane slicing," 2019[1] - In
    this paper, authors propose a CNN classifier base on image bit-plane
    slicing. The purpose is to improve recognition accuracy when we apply it to
    breast cancer image classification. Each texture image is decomposed into
    eight bit-plane images. Different bit-planes provide different levels and
    detail of image texture features. At the same time, they have also tested
    feature classification performance by each bit-plane respectively and the
    fusion of all bit-planes. CNN classifier is used for classification and
    recognition.
    Literature Review
    Bah and M. David, "Analysis of Breast Cancer Classification with Machine
    Learning based Algorithms," 2022[2] - In this paper, four Machine Learning
    algorithms, which are Random Forest, Support Vector Machine (SVM), K-
    Nearest Neighbors (KNN) and Convolutional Neural Networks (CNN), with
    five different breast cancer datasets are tested and analyzed to verify their
    performance in a binary classification of breast cancer. The results show that
    CNN obtained higher accuracy than the other tested algorithms in this type of
    data. This study will help future researchers in the breast cancer field to
    continue their research and focus on improving the performance of specific
    algorithms.
    Literature Review
    P. T. Nguyen, T. T. Nguyen, N. C. Nguyen, and T. T. Le, "Multiclass Breast
    Cancer Classification Using Convolutional Neural Network," 2019 - In this
    paper, the author uses CNN to classify and recognize breast cancer images from
    public Break His dataset. This dataset includes 7,909 breast cancer (BC)
    histopathology images with four benign subclasses and four malignant
    subclasses. Our new task with this dataset is the automated multi-classification
    of these breast cancer images in eight classes, which can help reduce death rates
    and save people's lives in the world. Our method is based on the resizing of
    original images for building the CNN model and classifying breast cancer
    classes.
    Problem Statement
    • Breast cancer is the most common cancer among women
    worldwide, and early detection is critical for successful treatment.
    In recent years, deep learning techniques have shown promising
    results in various medical imaging tasks, including breast cancer
    detection. Convolutional neural networks (CNN) are particularly
    effective in identifying patterns in medical images and can be used
    to classify mammograms as either benign or malignant.
    • To develop a deep learning model to classify cancer cells, thus
    facilitating doctors to achieve an accurate and fast diagnosis of
    cancer in an early stage.
    Objectives
    • This study aims to establish a deep learning model to classify
    cancer cells, thus facilitating doctors to diagnose breast cancer
    accurately and quickly in an early stage.
    • To improve the accuracy and speed of detection using deep
    learning techniques.
    • To deploy the application on the cloud, enabling anyone to access
    it via the internet at any time and from anywhere.
    Proposed System
    Proposed Work
    • The images have been resized to a consistent size (224x224) that suits the
    network architecture. The pixel values have been normalized to a suitable
    range, such as [0, 1].
    • Custom CNN architecture has been designed for the breast cancer detection
    task. The architecture has consisted of several convolutional layers followed
    by pooling layers, enabling the network to capture important features at
    various scales within the Histopathological images.
    • Fully connected layers have been added at the end of the network to classify
    the extracted features and make predictions. Techniques such as dropout and
    batch normalization have been employed to enhance the model's performance
    and mitigate overfitting.
    • These techniques have helped improve the model's accuracy and
    generalization ability in detecting breast cancer from the Histopathological
    images.
    Dataset used
    • The Breast Cancer Histopathological Image Classification (BreakHis) is
    composed of 9,109 microscopic images of breast tumor tissue collected from
    82 patients using different magnifying factors (40X, 100X, 200X, and 400X).
    • To date, it contains 2,480 benign and 5,429 malignant samples (700X460
    pixels, 3-channel RGB, 8-bit depth in each channel, PNG format).
    • This database has been built in collaboration with the P&D Laboratory –
    Pathological Anatomy and Cytopathology, Parana, Brazil.
    Algorithm used
    • custom CNN architecture has been designed for the breast cancer detection
    task.
    • The architecture has consisted of several convolutional layers followed by
    pooling layers, enabling the network to capture important features at various
    scales within the histopathological images.
    • Fully connected layers have been added at the end of the network to classify
    the extracted features and make predictions.
    • Techniques such as dropout and batch normalization have been employed to
    enhance the model's performance and mitigate overfitting.
    • These techniques have helped improve the model's accuracy and
    generalization ability in detecting breast cancer from the histopathological
    images.
    05/06/2023
    05/06/2023
    05/06/2023
    05/06/2023
    05/06/2023
    05/06/2023
    In the BreakHis dataset, which consists of
    histopathological images of breast cancer,
    several visual features can be used to identify
    malignant (cancerous) versus benign (non-
    cancerous) cases. These features are typically
    extracted from the images - morphological
    features such as cell shape, size, nuclear-to-
    cytoplasmic ratio, nuclear shape irregularities,
    chromatin texture, and presence of mitotic
    figures. Malignant cells often exhibit abnormal
    characteristics in comparison to benign cells.
    With the advent of deep learning, convolutional
    neural networks (CNNs) can be trained to
    extract relevant features automatically. These
    deep learning models can learn hierarchical
    representations of the input images, enabling
    them to identify patterns and subtle differences
    between benign and malignant cases.

    05/06/2023
    05/06/2023
    Conclusion
    CNNs have demonstrated exceptional accuracy in analyzing histopathological
    images, surpassing human performance in many cases. These deep learning
    models can effectively identify subtle patterns and features indicative of breast
    cancer, aiding in early and accurate detection. This is crucial because early
    detection significantly improves the chances of successful treatment and overall
    patient outcomes. The benefits of CNN-based breast cancer detection extend to
    doctors and medical professionals. By leveraging these advanced algorithms,
    doctors can benefit from a powerful tool that assists them in making more
    accurate and efficient diagnoses. The CNNs can process large volumes of
    histopathological images rapidly, allowing doctors to handle a larger number of
    cases and potentially reducing the burden on healthcare systems. Moreover,
    CNNs can provide a second opinion or act as a support system for doctors,
    enhancing their decision-making process and reducing the risk of misdiagnosis.
    Future Scope
    One aspect of future work involves developing CNN models that
    can effectively handle large-scale datasets of histopath images. As
    the volume of medical data continues to grow, it becomes crucial to
    train CNNs on diverse and extensive datasets to ensure robust
    performance. Additionally, researchers can explore transfer learning
    approaches, where pre-trained CNN models are fine-tuned using
    histopath images specific to breast cancer. This transfer learning
    strategy can leverage the knowledge learned from other related
    image recognition tasks, reducing the need for extensive data
    collection and annotation.
    References
    • https://www.who.int/news-room/fact-sheets/detail/breast-cancer

    • https://www.cancer.net/cancer-types/breast-cancer/statistics

    • A. Bah and M. Davud, "Analysis of Breast Cancer Classification with Machine Learning based Algorithms," 2022
    2nd International Conference on Computing and Machine Intelligence (ICMI), 2022, pp. 1-4.

    • Eldin, Sara & Hamdy, Jana & Adnan, Ganna & Hossam, Maysoon & Elmasry, Noha & Mohammed, Ammar. (2021).
    Deep Learning Approach for Breast Cancer Diagnosis from Microscopy Biopsy Images. 216-222.
    10.1109/MIUCC52538.2021.9447653.

    • Hasan, Asmiya & Bhoumik, Partho & Tauki, Md. Sadik Yasir & Chowdhury, Eshanee. (2022). Multiclass Breast
    Cancer Classification Using Convolutional Neural Networks. 10.13140/RG.2.2.11057.04963.

    • S. Ruban, M. Moosa Jabeer and R. Shenoy Basti, "Improvising Breast Cancer detection using CNN, VGG and SSD
    Algorithms," 2022 Third International Conference on Intelligent Computing Instrumentation and Control
    Technologies (ICICICT), Kannur, India, 2022, pp. 1772-1776, doi: 10.1109/ICICICT54557.2022.9918009.

    • S. N. Raju, K. Snehaja and B. Srinivas, "Ontology Based Disease Prediction System," 2021 6th International
    Conference on Communication and Electronics Systems (ICCES), Coimbatre, India, 2021, pp. 1-6, doi:
    10.1109/ICCES51350.2021.9489132.
    Thank You

    You might also like