International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 05 | May 2022
p-ISSN: 2395-0072
www.irjet.net
BREAST CANCER DETECTION USING MACHINE LEARNING Geetha P, Sneha R, Sneka.S, Subhiksha.R Electronics and Communication Engineering, Hindusthan College of Engineering And Technology, Coimbatore , Tamil Nadu , India -------------------------------------------------------------------------***----------------------------------------------------------------------ABSTRACT According to global statistics, breast cancer (BC) is one of the most frequent diseases among women globally, accounting for the majority of new cancer cases and cancer-related deaths, making it a serious public health issue in today's society. Early detection of BC improves the prognosis and chances of survival by allowing patients to receive timely clinical treatment. Patients may avoid unneeded therapies if benign tumours are classified more precisely. The classification method utilised in this paper is Modified CNN based feature extraction with transfer learning (ANN). A gaussian kernel approach is utilised for segmentation expectation maximisation .
INTRODUCTION Human cancer is a multifaceted illness characterised by genetic instability and the accumulation of numerous alternative molecules. Current diagnostic categories do not account for the whole clinical heterogeneity of malignancies and are insufficient to predict treatment efficacy and patient outcomes. The majority of currently used anti-cancer drugs do not distinguish between malignant and normal cells. Furthermore, cancer is frequently discovered and treated too late Cancer cells have already spread throughout the body. A large majority of individuals with breast, lung, colon, prostate, and ovarian cancer have hidden and over metastatic colonies at the time of clinical presentation. The effectiveness of therapy techniques is currently restricted. Use deep learning algorithms for breast cancer detection with mammography images in our planned effort.
© 2022, IRJET
|
Impact Factor value: 7.529
EXISTING METHODOLOGY Breast image preprocessing, mass detection, feature extraction, training data creation, and classifier training are the five phases in breast cancer detection covered in this work. To raise the contrast between the masses and the surrounding tissues, de-noising and boosting contrast techniques on the original mammography were used in the breast image pre-processing. After that, mass detection is used to locate the ROI. The ROI is then used to extract characteristics such as deep features, morphological features, texture features, and density features. The classifiers were trained with every image from the breast image dataset during the training phase, using their extracted features and labels. The welltrained classifier can so identify the mammography under diagnosis. A contrast enhancement method was utilised in this study to improve the contrast between the suspicious masses and the surrounding tissues. The fundamental concept is to make the original image's histogram equally distributed. The image's grey scale is expanded as a result of this procedure, which improves contrast and makes image details more visible.
|
ISO 9001:2008 Certified Journal
|
Page 399