Proc. SPIE. 10989, Big Data: Learning, Analytics, and Applications
KEYWORDS: Analytics, Image compression, Data modeling, Data storage, Magnetic resonance imaging, Image segmentation, Medical imaging, Machine learning, Functional magnetic resonance imaging, Evolutionary algorithms
Big data has been one of the hottest topics of scientific discussions in the recent years. In early 2000s, an industry analyst attempted to describe big data as the three Vs: Volume, Velocity, and Variability. With the new technologies such as Hadoop, it is now feasible to store and use extremely large volumes of data that comes in at an unprecedented velocity. The variability of this data can be large as it can come in different formats such as text documents, voice or video, and financial transactions. Big data analytics has been proven to be useful is various fields such as science, sports, advertising, health care, genomic sequence data, and medical imaging. This study presents a brief overview of big data analytics in medical imaging approaches with considering the importance of contemporary machine learning techniques such as deep learning.
The advancement of Internet of Things (IoT) technologies, such as low-cost embedded single board computers which integrate sensors, communication hardware, and processing power in one unit, has given more traction to the concept of Smart Cities. Having cheaper processing power at their disposal, the sensing units are capable of gathering increasingly larger amounts of raw data locally, which must be processed before being usable. One concern for this scheme is the amount of infrastructure and network bandwidth needed to transfer the data from the acquisition location to a server, which may be miles away, for further processing. The bandwidth available to the sensor network, distributed through the city, is expanding in a lower rate than the size and bandwidth demand of the network it serves. Therefore, transferring the unprocessed data to a central server does not seem feasible unless major compromises are made in terms of data resolution and size. This paper proposes a local big data based preprocessing scheme before the data is transferred to the storage. Using this scheme can free up the network bandwidth, exploit the otherwise wasted local processing power, and release processing load from the central server, allowing it to serve a larger network without the need for more powerful hardware. By making efficient use of network infrastructure the smart city applications are more affordable and scalable.
Reducing a graph model is extremely important for the dynamical analysis of large-scale networks. In order to approximate the behavior of such a system it is helpful to be able to simplify the model. In this paper, the graph reduction model is introduced. This method is based on removing edges that close independent cycles in the graph. We apply this novel model reduction paradigm to brain networks, and show the differences between the model approximation error for various brain network graphs ranging from those of healthy controls to those of Alzheimer's patients. The graph simplification for Alzheimer's brain networks yields the smallest approximation error, since the number of independent cycles is smaller than in either the healthy controls or mild cognitive impairment patients.
Leader-follower controllability in brain networks which are affected neurodegenerative diseases can provide important biomarkers relevant for disease evolution. The brain network is viewed as a dynamic system where the nodes interact via neighbor-based Laplacian feedback rules. The network has cooperative connections between the nodes described by positive weights along with competitive connections which are described by negative connection weights. The nodes take the role of either leaders or followers, thus forming a leader-follower signed dynamic graph network. The results of this analysis can be easily generalized on unsigned brain networks. We apply the leader-follower concept to structural and functional brain networks with neurodegenerative diseases (dementia) and show that the found leaders represent important biomarkers for disease evolution. In other words, the leader nodes drive the network towards deteriorating cognitive states.