Sharan Vaswani

Performance evaluation of medical imaging algorithms on Intel ® MIC platform

Heterogeneous computer architectures, where CPUs co-exist with accelerators such as vector coprocessors, GPUs and FPGAs, are rapidly evolving to be powerful platforms for tomorrows exa-scale computing. The Intel® Many Integrated Core (MIC) architecture is Intels first step towards heterogeneous computing. This paper investigates the performance of the MIC platform in the context of medical imaging and signal processing. Specifically, we analyze the achieved performance of two popular algorithms: Complex Finite Impulse Response (FIR) filtering which is used in ultrasound signal processing and Simultaneous Algebraic Reconstruction Technique (SART) which is used in 3D Computed tomography (CT) volume reconstruction. These algorithms are evaluated on Intel® Xeon Phi™ using Intels heterogeneous offload model. Our analysis indicates that execution times of both of these algorithms are dominated by the memory access times and hence effective cache utilization as well as vectorization play a significant role in determining the achieved performance. Overall, we perceive that Intel® MIC is an easy-to-program accelerator of the future that shows good potential in terms of performance.

Fast 3D Salient Region Detection in Medical Images Using GPUs

Automated detection of visually salient regions is an activearea of research in computer vision. Salient regions can serve as inputs for object detectors as well as inputs for region-based registration algorithms. In this paper, we consider the problem of speeding up computationally intensive bottom-up salient region detection in 3D medical volumes. The method uses the Kadir–Brady formulation of saliency. We show that in the vicinity of a salient region, entropy is a monotonically increasing function of the degree of overlap of a candidate window with the salient region. This allows us to initialize a sparse seed point grid as the set of tentative salient region centers and iteratively converge to the local entropy maxima, thereby reducing the computation complexity compared to the Kadir–Brady approach of performing this computation at every point in the image. We propose two different approaches for achieving this. The first approach involves evaluating entropy in the four quadrants around the seed point and iteratively moving in the direction that increases entropy. The second approach we propose makes use of mean shift tracking framework to affect entropy maximizing moves. Specifically, we propose the use of uniform pmf as the target distribution to seek high entropy regions. We demonstrate the use of our algorithm on medical volumes for left ventricle detection in PET images and tumor localization in brain MR sequences.

Modeling non-progressive phenomena for influence propagation

Most previous work on modeling influence propagation has focused on progressive models, i.e., once a node is influenced (active) the node stays in that state and cannot become inactive. However, this assumption is unrealistic in many settings where nodes can transition between active and inactive states. For instance, a user of a social network may stop using an app and become inactive, but again activate when instigated by a friend, or when the app adds a new feature or releases a new version. In this work, we study such non-progressive phenomena and propose an efficient model of influence propagation. Specifically, we model influence propagation as a continuous-time Markov process with 2 states: active and inactive. Such a model is both highly scalable (we evaluated on graphs with over 2 million nodes), 17-20 times faster, and more accurate for estimating the spread of influence, as compared with state-of-the-art progressive models for several applications where nodes may switch states.

Fast 3D structure localization in medical volumes using CUDA-enabled GPUs

Effective and fast localization of anatomical structures is a crucial first step towards automated analysis of medical volumes. In this paper, we propose an iterative approach for structure localization in medical volumes based on the adaptive bandwidth mean-shift algorithm for object detection (ABMSOD). We extend and tune the ABMSOD algorithm, originally used to detect 2D objects in non-medical images, to localize 3D anatomical structures in medical volumes. For fast localization, we design and develop optimized parallel implementations of the proposed algorithm on multi-cores using OpenMP, and on GPUs using CUDA. We evaluate the quality, performance and scalability of the proposed algorithm on Computed Tomography (CT) volumes for various structures.

Influence maximization in bandit and adaptive settings

The objective of viral marketing is to leverage a social network to spread awareness about a specific product in the market through information propagation via word-of-mouth. A closely related problem is that of influence maximization which aims to identify the ‘best’ set of ‘influential’ users (seeds) to give discounts or free products to, such that awareness about the product is maximized. We study two relatively unexplored variants of influence maximization (IM) in social networks. Conventional IM algorithms assume the structure of the social network and edge weights to be known. Though the structure of the network is usually known, edge weights need to be estimated from past data. In the first part of this thesis, we tackle the real but difficult problems of (i) learning these edge weights online and (ii) maximizing influence in the network when no past data is available as input. We adopt a combinatorial multiarmed bandit (CMAB) paradigm and formulate the above problems respectively as (i) network exploration, i.e. incrementally minimizing the error in learned edge weights and (ii) regret minimization i.e. minimizing the loss in influence from choosing suboptimal seed sets over multiple attempts. Most previous work on influence maximization in social networks is limited to the non-adaptive setting in which the marketer is supposed to select all of the seed users up front. A disadvantage of this setting is that she is forced to select all the seeds based solely on a diffusion model. If the selected seeds do not perform well, there is no opportunity to course-correct. A more practical setting is the adaptive setting in which the marketer initially selects a batch of users and observes how seeding those users leads to a diffusion of product adoptions. Based on this market feedback, she formulates a policy for choosing the remaining seeds. We study adaptive offline strategies for two problems: (a) MAXSPREAD given a budget on number of seeds and a time horizon, maximize the spread of influence and (b) MINTSS given a time horizon and an expected number of target users, minimize the number of required seeds.