CAREER: An Integrated Framework for Data-Adaptive Representations and Algorithms in Visual Computing

这是为期四年的持续奖的第一年资助。这个职业项目的目标是开发理论框架、算法和计算工具，以改进数据表示、分析和可视化。为此，我们建议为低级信号处理任务(例如，内插)和高级数据查询任务开发数据自适应算法。对于高层次的任务，我们建议发展对体数据的混合表示(多分辨率体积变形模型)，这将对检测和模拟都有用。建议的表示和算法将在3D MR数据(乳腺癌检测和手术计划)和地震数据(层序地层学衍生对象的检测)上进行测试。教育部分力求解决视觉计算领域学生有效教学和培训方面的一些重要挑战。这份职业计划的教育目标具体如下：(1)通过促进跨系教学活动，促进休斯顿大学的跨学科视觉计算教育；(Ii)提供视觉计算教育的实践者视角，并解决视觉计算课程中的“现实世界”问题；以及(Iv)吸引新生进入视觉计算，并激励高中生对科学感兴趣。一个关于微创心脏手术模拟器原型的项目将被用作视觉计算跨学科教学的示范范例。领域专业知识和领域数据将通过与M.D.安德森癌症中心和Veritas探索服务公司的密切合作提供。

This is the first year funding of a four-year continuing award. The goal of this CAREER project is to develop the theoretical framework, algorithms, and computational tools for improved data representation, analysis, and visualization. To that end, we propose to develop data-adaptive algorithms for both low-level signal processing tasks (e.g., interpolation) and high-level data interrogation tasks. For the high-level tasks, we propose to develop hybrid representations for volumetric data (multiresolution volumetric deformable models) that will be useful for both detection and simulation. The proposed representations and algorithms will be tested on 3D MR data (breast cancer detection and surgery planning) and seismic data (detection of sequence stratigraphy derived objects). The educational component seeks to address some important challenges in the effective teaching and training of students in the Visual Computing field. In particular, the educational goals of this CAREER plan are the following: (1) to promote an interdisciplinary Visual Computing education at the University of Houston by fostering cross-departmental teaching activities, (ii) to provide a practitioner's perspective in Visual Computing education and to address "real-world" problems in the Visual Computing Curriculum, and (iv) to attract incoming students into Visual Computing and to motivate high-school students towards the sciences. A project on a prototype minimal invasive heart surgery simulator will be used as a demonstration paradigm for the interdisciplinary teaching of Visual Computing. Domain expertise and domain data will be provided through close collaborations with the M.D. Anderson Cancer Center and Veritas Exploration Services.