Energy Awareness training ppt for manufacturing process.pptx
PPT_ZEROTH_REVIEW.pptx
1. GALGOTIAS COLLEGE OF ENGINEERING AND
TECHNOLOGY
ELECTRONICS AND COMMUNICATION DEPARTMENT
BREAST CANCER DETECTION USING DEEP LEARNING
PROJECT GUIDE:
MR. DHINAKARAN M.
GROUP MEMBERS:
1) Ritik Lahoti (1900970310129)
2) Riyanshi Singh(1900970310130)
3) Saransh Chauhan(1900970310140)
4) Tanvi Sharma (1900970310171)
3. INTRODUCTION
• Image processing algorithms are commonly employed in various medical domains to increase disease detection
capability.
• Cancer is one of the most dangerous diseases, and early detection is a challenging task in medicine. The cancer
detection method entails classifying the image biopsy as malignant or benign.
• The project is based on breast cancer samples. After a comparative examination of commonly used methods in each
category, an appropriate and efficient approach is adopted in each of the design processes of the proposed framework
• Mammography, ultrasonography, and thermography are the three most common imaging methods utilized for the
screening to detect this condition.
• Mammography is ineffective for solid breasts; diagnostic ultrasound techniques are frequently used. In light of these
concerns, radiographic radiation can skip small tumors, and thermography may be a better method of cancer diagnosis
than ultrasound for smaller tumors
• The dataset which will be used in this project is microscopic biopsy images of breast cancer and microscopic biopsy
images of non-cancerous cells derived from the Breast Cancer Histopathological Database (BreakHis) and kaggle.
• We will implement the model with Machine learning , image processing and CNN.
4. OBJECTIVE
• To classify Cancerous and Non-cancerous cells using Image Processing and Machine Learning
Techniques.
• To find out the accuracy of the classification with image processing and data segmentation.
• To evaluate the efficacy of image enhancement in identifying breast cancer.
5. SYSTEM DESIGN
Microscopic Biopsy Images
Train and Evaluation Split
Build Model 2
Build Model 1
GlobalAveragePooling 2D
DenseNet 201
Maxpooling 2D
DenseNet 201
Classification
Prediction
Training and Evaluation
Benign Malignant
6. PARAMETERS TO BE ANALYSED
parameters to be analysed includes :
• Weights
• Biases
8. ACTION PLAN
AUG-SEP
Review of Literature
OCT-NOV
Feasibility study +
Requirement analysis of
Problem Domain
DEC-JAN
Data analysis,visualisation
for benign and malignant
dataset
FEB-MAR
Validation of Results
APR-MAY
Report Preparation