DETECTION OF DIFFERENT TYPES OF CANCER USING DEEP LEARNING

Abstract

We provide an application of deep learning for cancer detection. We firstly provide what is a cancer, statistics of it and if not detected and diagnosed early can lead to death and then we proposed different models like CNN and how it can helps to detect or classify a cancer wheather a diagnosing person has cancer or not from a medical image(dataset). Image processing techniques have lately been popular in a number of medical sectors for image improvement at early diagnostic and treatment stages, especially in cancer tumours like lung cancer and breast cancer, where the time factor is crucial in detecting anomalies in target photographs. We have collected the dataset from kaggle which is a collection of dataset website and the dataset has been preprocessed to remove artifacts like blur, noise etc by various data cleaning technique and has been trained on a Deep learning models like basic CNN architecture, AlexNet, VGG16, ResNet-50, basic CNN architecture with dilation rate 2.0 and concatenation of two different architecture. At last we evaluate and compare the model in terms of accuracy, precision, recall and F-Score using Confusion matrix.