Chapter 11 - Computational intelligence on medical imaging with artificial neural networks

Abstract

Listen

Translate article

Computer-aided diagnosis systems based on artificial intelligence is widely utilized recently by clinicians, particularly on medical imaging. In artificial intelligence, artificial neural networks are among the most preferred topics on imaging such as magnetic resonance imaging, computer tomography, X-ray, ultrasound etc. Recently, artificial neural networks are preferred by biomedical researchers owing to their automatic learning features capability. Artificial neural networks also developed over time and became the basis for deep learning and convolutional neural networks. Based on these algorithms, researchers have focused on developing the most accurate and simplest systems for computer aided on medical imaging and presented a rather wide range in the last few years. The aim of using artificial neural networks approaches is to recognize various disorders using image segmentation, image classification, image estimation, and extraction of features from images in the medical field, easily. Hence, these approaches help rather to clinicians, doctors, and patients with so many applications. In this chapter, firstly, based on artificial neural networks algorithms: deep neural networks, convolutional neural networks, recurrent neural networks, generative adversarial networks, autoencoders, etc. are introduced and then imaging types are briefly mentioned as well as the imaging used for which including some diseases. Therefore, some diseases (cancer, tumor, cysts, neural disorders: Alzheimer's disease, Parkinson's disease, pneumonia, infections, skin disorders, etc.) are also reviewed in this chapter. Moreover, these comprehensive reviews refer to the imaging datasets that are utilized in research and applied techniques, as well. This chapter presents current studies, improvements, problems, and suggestions that will shed light on future studies.