We rst develop a grouping mechanism that organizes individual edges into long contours, by encoding Gestalt factors of proximity, continuity, collinearity, and closure in a graph.
We typically recognize object shape as a whole - the entire geometric configuration of image tokens and the context they are in. We propose a linear programming LP formulation for effciently searching over exponentially many contourconfigurations. Used as integral shape matching units, they are powerful for preventing accidental alignment to isolated edges, dramatically reducing false shape detections in clutter.
These platforms will be implemented to detect and analyze chromosomal imbalances for whole genome array comparative genomic hybridization aCGH and single nucleotide polymorphisms SNP microarray-based technologies.
We have designed a customized high-resolution genome-wide microarray to detect genetic aberrations from congenital diaphragmatic hernia CDH patient samples. The incidence of CDH is approximately 1 in 2, newborns.
A formulation by semidefinite program Qihui zhu thesis SDP provides an efficient computational solution to this NP-hard problem, and the flexibility of expressing various bottom-up grouping cues.
CDH is a defect in the diaphragm that allows the stomach and intestines to move into the chest cavity and cause underdevelopment of the lungs. We then propose a set-to-set shape matching paradigm that measures and compares holistic shape configurations.
Computationally, set-to-set matching is a hard combinatorial problem. Detecting shape as a global pattern involves two key issues: The contours are characterized by their topologically ordered 1D structures, against otherwise chaotic 2D image clutter.
Contours or regions that are grouped from bottom-up often appear as half complete shapes that are easily recognizable. On the image side, we extend contour packing to regions, which has a fundamentally different topology. We also develop a primal-dual packing algorithm to quickly bound and prune solutions without actually running the LPs.
Bipartite graph packing is designed to cope with this change. The main challenge arises from the representation gap between image and model: About Qihui Zhu, Ph.
The holistic configuration is captured by shape features with a large spatial extent, and the long-range contextual relationships among contours. I am currently working on several research projects in the Dr.
Humans have an amazing ability to localize and recognize object shapes from nat- ural images with various complexities, such as low contrast, overwhelming background clutter, large shape deformation and significant occlusion. This thesis presents Contour Packing, a novel framework that detects shapes in a global and integral way, effectively bridging this representation gap.
Representing both the model and the image as a set of contours, we seek packing a subset of image contours into a complete shape formed by model contours.
A proper model captures long range geometric constraints among image tokens. Publicly Accessible Penn Dissertations. For best results, right-click and select "save as The unique feature of this approach is the ability to overcome unpredictable contour fragmentations.
Another project is to design and optimize genomic structural variation analysis platforms. One of the projects is to identify and characterize genetic aberrations associated with congenital diaphragmatic hernia CDH.
We are currently running this array platform on the CDH patients and control samples, and analyzing the data. Finally, we generalize set-to-set shape matching on more sophisticated structures arising from both the model and the image.
More than 80 percent of the patients are not associated with any known genetic syndrome, and the underlying mechanisms for their conditions remain unknown.
On the model side, we enrich the representation by compactly encoding part conguration selection in a tree, making holistic matching applicable to articulated objects.This thesis presents Contour Packing, a novel framework that detects shapes in a global and integral way, effectively bridging this representation gap.
Zhu, Qihui, "Shape Detection by Packing Contours and Regions" (). Publicly Accessible Penn Dissertations. iii THE IMPACT OF OIL RESERVES ON CHINA’S INFRASTRUCTURE INVESTMENT IN AFRICA Leixin Zhu, B.A. Thesis Advisor: Donna Morrison, Ph.D. ABSTRACT I investigate the empirical determinants of China’s infrastructure investment in Africa.
Shape Detection by Packing Contours and Regions Qihui Zhu University of Pennsylvania, This thesis presents Contour Packing, a novel framework that detects shapes in a global and integral way, SHAPE DETECTION BY PACKING CONTOURS AND REGIONS Qihui Zhu.
Zhu T., (), “Finite-Difference Modeling in 2-D Viscoelastic-Fractured media”: International Geophysical and Geological Conference and Exhibition SEG/EAGO Tyumen. SHAPE DETECTION BY PACKING CONTOURS Qihui Zhu A DISSERTATION in Computer and Information Science Presented to the Faculties of the University of Pennsylvania in Partial.
Qihui Zhu, Ph.D., is trying to identify and characterize genetic aberrations associated with congenital diaphragmatic hernia.Download