SILO



The weekly SILO Seminar Series is made possible through the generous support of the 3M Company and its Advanced Technology Group

3M

with additional support from the Analytics Group of the Northwestern Mutual Life Insurance Company

Northwestern Mutual

Cooperative Communication: Backpressure Algorithm with Mutual Information Accumulation | Online robust PCA and examples in computer vision

Nick Yanpei Liu and Laura Balzano, Graduate students in ECE

Date and Time: Nov 02, 2011 (12:30 PM)
Location: Orchard room (3280) at the Wisconsin Institute for Discovery Building

Abstract:

Cooperative Communication: Backpressure Algorithm with Mutual Information Accumulation
By Nick Yanpei Liu

We develop scheduling policies that maximize the stability region of a wireless network under the assumption that mutual information accumulation is implemented at the physical layer. When the link quality between nodes is not sufficiently high that a packet can be decoded within a single slot, the system can accumulate information across multiple slots, eventually decoding the packet. The result is an expanded stability region. We propose two dynamic scheduling algorithms. The first performs scheduling every Tslots, and approaches the boundary of the stability region as T gets large, but at the cost of increased average delay. The second introduces virtual queues for each link and constructs a virtual system wherein two virtual nodes are introduced for each link. The constructed virtual system is shown to have the same stability region as the original system.

Online robust PCA and examples in computer vision
By Laura Balzano

Low-dimensional linear subspaces are used to model the background of video for object tracking in computer vision. Recently developed algorithms for Robust PCA have been applied to the problem of separating out background from moving foreground objects in the video, and they achieve this goal with very high quality. However, the algorithms work in batch on several video frames at once. In this talk I will present GRASTA (Grassmannian Robust Adaptive Subspace Tracking Algorithm), an online stochastic gradient algorithm for Robust PCA. GRASTA is uniquely suited to the video application, because it operates one frame at a time but still achieves very high quality separation. GRASTA can even identify the background when many of the pixels are discarded, making separation even faster. On some popular video benchmarks, GRASTA can separate background from foreground in 30-60 frames per second on my laptop in Matlab. I will discuss the algorithm, its performance, and some remaining challenges in the computer vision application.