A Study On Various Segmentation And Machine Learning Algorithms Used For Early Detection Of Cancer.

Prasenjit Dhar,K. Suganya Devi,Satish Kumar Satti,P. Srinivasan

Cancer is the most dangerous disease. Diagnosing the cancer cell in early stages plays a vital role in saving human life and for successful treatment. Spreading of cancer cells to different parts of the body can be stopped by removing the benign cells (cancer cells in the early stage). Whereas, malignant tumor (cancer cells in late stages) which are having aggressive spreading capacity, affects different other parts of the body and cannot be controlled. This paper presents a study on five different types of cancer viz., breast, brain, lung, liver and skin; and different published techniques on detecting these cancers, which help the researcher to understand the current techniques and aids to develop new structure that gives better and accurate result. It also focus on different segmentation (ACM, PSO, UNet, watershed etc), cancer feature extraction, cancer feature reduction (PCA, LDA, SVD). Also it discuss different cancer classification using machine learning and clustering (SVM, KNN, Bayesian, Neuro fuzzy, k-mean algorithm, GANs etc.), deep learning (CNN, ResNet, VGG etc) technique, also discuss about different evaluating method.

Volume 12 | Issue 6

Pages: 1416-1432

DOI: 10.5373/JARDCS/V12I2/S20201339