Convolutional neural networks for computer-aided detection or diagnosis in medical image analysis: An overview
146
Citation
120
Reference
10
Related Paper
Citation Trend
Abstract:
Computer-aided detection or diagnosis (CAD) has been a promising area of research over the last two decades. Medical image analysis aims to provide a more efficient diagnostic and treatment process for the radiologists and clinicians. However, with the development of science and technology, data interpretation manually in the conventional CAD systems has gradually become a challenging task. Deep learning methods, especially convolutional neural networks (CNNs), are successfully used as tools to solve this problem. This includes applications such as breast cancer diagnosis, lung nodule detection and prostate cancer localization. In this overview, the current state-of-the-art medical image analysis techniques in CAD research are presented, which focus on the convolutional neural network (CNN) based methods. The commonly used medical image databases in literature are also listed. It is anticipated that this paper can provide researchers in radiomics, precision medicine, and imaging grouping with a systematic picture of the CNN-based methods used in CAD research.Keywords:
Computer-Aided Diagnosis
Cite
Citations (5)
Computer-Aided Diagnosis
Virtual colonoscopy
Cite
Citations (0)
Computer aided diagnosis (CAD) is a currently applied computerized analysis of medical images and is widely used as a tool in detection and differential diagnosis of abnormalities in medical imaging. In clinical practice CAD became a very useful implement in various imaging techniques such as mammography, radiography, CT, MRI or diagnostic ultrasound. CAD became also a major research subject in clinical radiology e.g. in detection of pulmonary nodules in chest CT or breast cancer on mammograms. CAD concept, unlike known from 1960s automated computer diagnosis concept (ACD), which is based on computer algorithms only takes into account equally the roles of physician and computer. Current directions in CAD development focus on continuous improvement of overall performance in detection of pathological lesions and assembling CAD schemes as packages implemented as a part of PACS system. The aim of this article is presentation of the concept of computer aided diagnosis, current opportunities of adopting CAD systems into various imaging techniques and new directions in CAD development.
Computer-Aided Diagnosis
Presentation (obstetrics)
Cite
Citations (6)
Computer-aided diagnosis (CAD) systems have been developed in many institutions. CAD is able assist radiologists and physicians in detecting lesions and in differentiating benign from malignant lesions on medical images. The output from CAD can be used as a "second opinion" to assist radiologists in their interpretations and improve diagnostic accuracy. It is important to note that CAD is not designed for the automated diagnostic system. A few commercial CAD systems have recently been developed by venture companies, with pre-marketing approval having been obtained from the U.S. Food and Drug Administration (FDA). It is clear that CAD has practical value in radiologic diagnosis. In this paper, the background of CAD research and important topics, including the usage of an image database and methods of evaluating the performance of typical CAD schemes, are discussed. The number of CAD researchers in Japan is not yet sufficient. Radiological technologists who have a strong interest in CAD research will be good partners for radiologists in developing CAD schemes.
Computer-Aided Diagnosis
Cite
Citations (1)
Abstract Computer‐aided diagnosis (CAD) is defined as a diagnosis made by a radiologist who takes into account the output of a computerized analysis of a medical image. There are in general two different types of CAD schemes: one to locate suspicious regions in the image, computer‐aided detection (CADe) and another to classify a suspicious region into one of two or more classes, computer‐aided diagnosis (CADx). The goal of CAD is to improve the performance of radiologists, to reduce variation within and between radiologists, and to make radiologists more efficient. This article describes motivation for developing CAD, the basic principles, current issues in CAD research, and future areas for development.
Computer-Aided Diagnosis
Computer-aided
Cite
Citations (2)
Deep learning has made spectacular achievements in analysing natural images, but it faces challenges for medical applications partly due to inadequate images.Aiming to classify malignant and benign pulmonary nodules using CT images, we explore different strategies to utilize the state-of-the-art deep convolutional neural networks (CNN).Experiments are conducted using the Lung Image Database Consortium image collection (LIDC-IDRI), which is a public database containing 1018 cases. Three strategies are implemented including to 1) modify some state-of-the-art CNN architectures, 2) integrate different CNNs and 3) adopt transfer learning. Totally, 11 deep CNN models are compared using the same dataset.Study demonstrates that, for the model modification scheme, a concise CifarNet performs better than the other modified CNNs with more complex architectures, achieving an area under ROC curve of AUC = 0.90. Integrated CNN models do not significantly improve the classification performance, but the model complexity is reduced. Transfer learning outperforms the other two schemes and ResNet with fine-tuning leads to the best performance with an AUC = 0.94, as well as the sensitivity of 91% and an overall accuracy of 88%.Model modification, model integration, and transfer learning can play important roles to identify and generate optimal deep CNN models in classifying pulmonary nodules based on CT images efficiently. Transfer learning is preferred when applying deep learning to medical imaging applications.
Transfer of learning
Cite
Citations (54)
This chapter contains sections titled: Computer-Aided Diagnosis (CAD) Computer-Aided Detection and Diagnosis (CAD) without PACS Conceptual Methods of Integrating CAD with DICOM PACS and MIII Computer-Aided Detection of Small Acute Intracranial Hemorrhage (AIH) on Computer Tomography (CT) of Brain Evaluation of the CAD for AIH From System Evaluation to Clinical Practice Summary of Computer-Assisted Diagnosis Reference
Computer-Aided Diagnosis
DICOM
Computer-aided
Cite
Citations (0)
Manual Fruit classification is the traditional way of classifying fruits. It is manual contact-labor that is time-consuming and often results in lesser productivity, inconsistency, and sometimes damaging the fruits (Prabha & Kumar, 2012). Thus, new technologies such as deep learning paved the way for a faster and more efficient method of fruit classification (Faridi & Aboonajmi, 2017). A deep convolutional neural network, or deep learning, is a machine learning algorithm that contains several layers of neural networks stacked together to create a more complex model capable of solving complex problems. The utilization of state-of-the-art pre-trained deep learning models such as AlexNet, GoogLeNet, and ResNet-50 was widely used. However, such models were not explicitly trained for fruit classification (Dyrmann, Karstoft, & Midtiby, 2016). The study aimed to create a new deep convolutional neural network and compared its performance to fine-tuned models based on accuracy, precision, sensitivity, and specificity.
Deep Neural Networks
Cite
Citations (2)
Deep learning is now an active research area. Deep learning has done a success in computer vision and image recognition. It is a subset of the Machine Learning. In Deep learning, Convolutional Neural Network (CNN) is popular deep neural network approach. In this paper, we have addressed that how to extract useful leaf features automatically from the leaf dataset through Convolutional Neural Networks (CNN) using Deep Learning. In this paper, we have shown that the accuracy obtained by CNN approach is efficient when compared to accuracy obtained by the traditional neural network.
Deep Neural Networks
Cite
Citations (7)
Because of the difficulty of CT imaging diagnosis of hepatocellular carcinoma, it is necessary to research the system of Computer-aided Diagnosis (CAD). Most practical diagnostic schemes use single image features to input single-phase images for classification. In this paper, a diagnosis system is designed based on BP neural network after dimensionality reduction by PCA, which combines multiple features and multi-phase information. The experimental results show that the comprehensive accuracy of this system reaches 96.98%, which is superior to the single-phase CAD system and comparable with conventional CNN.
Computer-Aided Diagnosis
Liver Cancer
Cite
Citations (4)