Detection and Classification of Multi Cancers using Transfer Learning

Worldwide, cancer is regarded as the leading cause of death; brain and lung cancers are especially difficult to identify and categorize in their early stages. For effective cancer treatment and better patient outcomes, early cancer detection and classification are essential. However, because these cancers are complex and don’t always exhibit distinct symptoms, conventional diagnostic techniques frequently fall short of identifying them in their early stages. Deep learning has demonstrated significant impact in the identification and categorization of a number of illnesses, including cancer. But there is still space for improvement in the precision of current deep learning models for the identification and categorization of early-stage cancers, especially when it comes to malignancies like brain and lung cancers. The process of creating precise methodologies is fraught with difficulties. This study suggests using deep learning based transfer learning models to classify images of various cancer types, such as lung, brain, Acute Lymphoblastic Leukemia and oral cancer. In this work, Transfer learning model is evaluated against the classification of images exhibiting malignant features. Pre-trained CNN models include ResNet50, MobileNet, DenseNet and VGG, use the knowledge acquired from the ImageNet dataset to identify various types of cancer. The utilization of the ResNet50 model within this research has yielded notably superior results. These outcomes are crucial for making precise decisions in the medical domain.