There is no doubt that medical imaging has become indispensable in disease diagnosis and therapy. With advances in medical imaging, new imaging modalities and methodologies such as cone-beam/multi-slice CT, 3D ultrasound imaging, tomosynthesis, diffusion-weighted magnetic resonance imaging (MRI), positron-emission tomography (PET)/CT, electrical impedance tomography, and diffuse optical tomography, the imaging information available for clinical decision making has been erupting. To take full advantage of medical imaging, new algorithms and methods are demanded in the medical imaging field to better process the overwhelming volume of data. However, due to the large variations and complexity in the medical imaging data, it is generally hard to derive analytic solutions or simple equations to represent objects such as lesions and anatomies in medical imaging data. Therefore, tasks in medical imaging require “learning from examples” for accurate representation of data and prior knowledge, which is exactly the focus of machine learning. Thus, machine learning in medical imaging has become one of the most promising growing fields. Machine learning is currently playing an essential role in the medical imaging field, including computer-aided diagnosis, image segmentation, image registration, image fusion, image-guided therapy, image annotation, and image database retrieval.

The main aim of this special issue is to help advance the scientific research within the broad field of machine learning in medical imaging. The special issue was planned in conjunction with the International Workshop on Machine Learning in Medical Imaging 2011 [1]. Being the first workshop on this topic, it has been successfully held together with the International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) for four consecutive years since 2010. This special issue is one in a series of special issues of journals on this topic [2]. The focus of the special issue is on major trends and challenges in this area. It presents work aimed at identifying new cutting-edge techniques and their use in medical imaging.

The majority of the papers included in this special issue were presented in a preliminary form at the International Workshop on Machine Learning in Medical Imaging, held in Toronto, Canada on September 18th, 2011. Together with some other new submissions, the submitted papers underwent two rounds of rigorous peer review. At the end, a small set of papers in total of 11 was selected to be included in this special issue to ensure a high quality outcome. There were quite several other good papers, which however, were not able to be included into the special issue due to the limited space. Those papers were later recommended for publication as regular papers of the journal, Machine Vision and Applications.

The special issue covers a wide spectrum of machine learning methods with their applications in different medical imaging modalities and targeted organs. The biomedical imaging modalities included in this special issue include CT [3, 5, 10], PET [4], MRI [6, 11–13], histology images [7, 9], and echocardiograms [8]. The targeted organs or structures include vertebrae [3, 10], thoracic structures [5], breast tissue [6, 9], lung [7], heart [8], knees [11, 13], brain [12], as well as a phantom [4]. The fusion of machine learning methods and medical imaging applications further demonstrates the versatility and productivity of this multidisciplinary research after years of development. In this special issue, a dictionary-based classification method using a cascade of simultaneous orthogonal matching pursuit classifiers [3] was applied to image-guided spine surgery. The statistical lesion activity computation (SLAC) [4], topological characterization of lesion enhancement patterns [6], a morphological tissue pattern-based classification [7], and the random subspace classifier ensembles [9] were used for cancer detection and classification. An online boosting algorithm [5] was designed for automatically detecting anatomic structures. As for the classical task of medical image segmentation, new attempts including using AdaBoost for defining local and global priors [8], sample-driven active shape models [10], unified extreme learning machines and discriminative random fields [13] were investigated. A partial least square-based method was developed for diagnosing knee osteoarthritis [11]. A cortical thickness network was established using the kernel-based method for predicting Alzheimer’s disease [12].

We would like to thank all the authors for their excellent contributions to this special issue and to all the reviewers for their high quality reviews and constructive suggestions. We hope that this special issue will inspire further ideas for creative research, advance the field of machine learning in medical imaging, and facilitate the translation of the research from bench to bedside.