EI Compendex Source List（2022年1月）
EI Compendex Source List（2020年1月）
EI Compendex Source List（2019年5月）
EI Compendex Source List（2018年9月）
EI Compendex Source List（2018年5月）
EI Compendex Source List（2018年1月）
中国科学引文数据库来源期刊列
CSSCI(2017-2018)及扩展期刊目录
2017年4月7日EI检索目录（最新）
2017年3月EI检索目录
最新公布北大中文核心期刊目录
SCI期刊（含影响因子）
论文范文
1. Introduction With increasing incidence and mortality, cancer has always been a leading cause of death for many years. According to American Cancer Society, there are around 1,685,210 new cases and 595,690 deaths in 2016 [1]. It was reported that the 5-year survival rate for the cancer patients diagnosed in early stage was as high as 90% [2]. In this regard, early and precise diagnosis is critical for better prognosis of cancer. Molecular imaging is an imaging technique to visualize, characterize, and measure biological procedures at molecular and/or cellular level [3] and has been considered as a powerful tool for early detection of cancer. Compared with anatomical imaging techniques, molecular imaging is more promising in diagnosing cancer in the early stage, as it is capable of signaling the molecular or physiological alterations in cancer patients which may happen before the obvious anatomical changes. Molecular imaging is also helpful in individualized therapy as it can reflect the treatment response at the molecular level. Therefore, molecular imaging has been widely used in cancer management. The current molecular imaging modalities in clinical practice include contrast-enhanced computed tomography (CT), contrast-enhanced magnetic resonance (MR) imaging, MR spectroscopy, and nuclear medicine such as single photon emission computed tomography (SPECT) and positron emission tomography (PET). Visual assessment conducted by the radiologists is the most common way to analyze these images. However, subtle changes in molecular images may be difficult to detect by visual inspection as the target-to-background ratio in these images is not that significant. In addition, visual interpretation by clinicians not only is time-consuming but also usually causes large variations across interpreters due to the different experience. The emerging intelligent techniques are of great potential in solving these problems by making the image interpretation automated. Machine learning-based image processing has been widely used in the domain of medical imaging analysis. Conventional machine learning techniques require the artificial intervention of feature extraction and selection and thus are still somehow subjective. In addition, the subtle and distributed changes may be ignored with artificial feature calculation and selection. Fully automated techniques are expected to integrate the local and global information for more accurate interpretation. Deep learning as a state-of-the-art machine learning technique may solve the challenges aforementioned by abstracting higher level features and improving the predictions from data with deep and complex neural network structures [4]. 1.1. Deep Learning The deep architectures and algorithms have been summarized [5, 6]. Compared with the conventional machine learning techniques, deep learning has shown some advantages [5, 6]. First, deep learning can automatically acquire much richer information in a data-driven manner and these features are usually more discriminative than the traditional hand-crafted features. Second, deep learning models are usually trained in an end-to-end way; thus the feature extraction, feature selection, and classification can be conducted and gradually improved through supervised learning in an interactive manner [7]. Therefore, deep learning is promising in a wide variety of applications including cancer detection and prediction based on molecular imaging, such as in brain tumor segmentation [8], tumor classification, and survival prediction. Deep learning-based automated analysis tools can greatly alleviate the heavy workload of radiologists and physicians caused by the popularity of molecular imaging in early diagnosis of cancer as well as enhance the diagnostic accuracy, especially when there exist subtle pathological changes that cannot be detected by visual assessment. 
|
|
|
