ABSTRACT
In the modern world, cancer is one of the most infectious illnesses. With artificial food and lifestyle induced in the modern lifestyle and more exposure to carcinogenic substances, more individuals are getting one of different sorts of cancer. Early detection of cancer can help in faster and efficient recovery and thus different detection techniques have evolved for the purpose. Imaging-based diagnosis is a well-established technique and the manual process requires a human expert to examine these reports. In this chapter, we have presented a comprehensive analysis of deep learning-based models and architectures that can be implemented by processing image scans to detect and identify various types of cancer. These models are very fast and more than 95% accurate and save a lot of human time with better accuracy. We have also compiled a list of publically available data sets that will help in training and testing such systems. In the end, we include the performance analysis of various models on different cancer types and conclude that approaches based on deep learning are certainly much faster and better than the traditional cancer detection and classification methods.
