John Recker

Perturbative time and frequency allocations for RFID reader networks

RFID reader networks often have to operate in frequency and time constrained regimes. One approach to the allocation of frequency and time to various readers in such regimes is to perturb the network slightly so as to ease the constraints. We investigate how to perform these perturbations in a manner that is profitable from time and frequency allocation point of view.

A parallel error diffusion implementation on a GPU

In this paper, we investigate the suitability of the GPU for a parallel implementation of the pinwheel error diffusion. We demonstrate a high-performance GPU implementation by efficiently parallelizing and unrolling the image processing algorithm. Our GPU implementation achieves a 10 - 30x speedup over a two-threaded CPU error diffusion implementation with comparable image quality. We have conducted experiments to study the performance and quality tradeoffs for differences in image block sizes. We also present a performance analysis at assembly level to understand the performance bottlenecks.

Plane-dependent error diffusion on a GPU

In this paper, we study a plane-dependent technique that reduces dot-on-dot printing in color images, and apply this technique to a GPU-based error diffusion halftoning algorithm. We design image quality metrics to preserve mean color and minimize colorant overlaps. We further use randomized intra-plane error filter weights to break periodic structures. Our GPU implementation achieves a processing speed of 200 MegaPixels/second for RGB color images, and a speedup of 30 - 37x over a multi-threaded implementation on a dual-core CPU. Since the GPU implementation is memory bound, we essentially get the image quality benefits for free by adding arithmetic complexities for inter-plane dependency and error filter weights randomization.

Optimal scheduling for networks of RFID readers

Devising switching schemes for networks of colliding and correlated RFID readers is a core challenge in the deployment of RFID networks. We derive optimal scheduling schemes for readers in RFID networks in four cases of practical importance. Most other cases can be reduced to a combination of these basic cases.

Perceiving Gloss in Surfaces and Images

Color Appearance Models are successfully used to model the color perception differences seen when the same stimuli are presented on different media, e.g. hard copy or a self-luminous display. It is currently unknown if the similar effects are present in gloss perception and if there is need for Gloss Appearance Models. Gloss communication, and the higher level material appearance communication is becoming more important everyday with the increase in customized manufacturing and the need for the costumer to preview a final product while short-runs, time and cost constraints prohibit the use of hard-copy proofs. Three experiments are proposed in order to analyze this phenomenon. The Gloss matching performance of observers on real objects is first going to be studied. Then, the same experiment will be repeated with synthetic images. Finally, a cross-media matching experiment will be performed, where the observers will have to match a real material with synthetic representations. The same trend was observed in the experiment using only real objects and in the cross-media situation, where a high matching accuracy was obtained for low gloss samples, and the gloss of mid and high gloss samples was underestimated. The same accuracy for low gloss samples was obtained for the experiment with only synthetic images, but mid and high gloss samples were overestimated. The sensitivity of the observers was higher when only real samples were used, it decreased when the display was used due the lack of visual disparity and multiple viewing conditions, and it was lowest on the last experiment, influenced by the multiple media and the above limitations.

Font rendering on a GPU-based raster image processor

Historically, in the 35 years of digital printing research, raster image processing has always lagged behind marking engine technology, i.e., we have never been able to deliver rendered digital pages as fast as digital print engines can consume them. This trend has resulted in products based on throttled digital printers or expensive raster image processors (RIP) with hardware acceleration. The current trend in computer software architecture is to leverage graphic processing units (GPU) for computing tasks whenever appropriate. We discuss the issues for rendering fonts on such an architecture and present an implementation.

Evaluating a Personal Communication Tool: Sidebar

By more closely integrating email with the web we aim to bring organization to email and more collaboration to the web. To this end we developed the Sidebar, a web-browser plug which displays email messages which link to the currently displayed URL. We conducted longitudinal studies on two versions of Sidebar to observe the usage of Sidebar and determine if it improves communications productivity. We found that providing an email summary in Sidebar resulted in raised awareness of the email collaborations, increased serendipitous discovery of information, and resulted in higher reported communication productivity. This paper summarizes Sidebars operation, describes the user studies, and presents conclusions.

A GPU accelerated PDF transparency engine

As commercial printing presses become faster, cheaper and more efficient, so too must the Raster Image Processors (RIP) that prepare data for them to print. Digital press RIPs, however, have been challenged to on the one hand meet the ever increasing print performance of the latest digital presses, and on the other hand process increasingly complex documents with transparent layers and embedded ICC profiles. This paper explores the challenges encountered when implementing a GPU accelerated driver for the open source Ghostscript Adobe PostScript and PDF language interpreter targeted at accelerating PDF transparency for high speed commercial presses. It further describes our solution, including an image memory manager for tiling input and output images and documents, a PDF compatible multiple image layer blending engine, and a GPU accelerated ICC v4 compatible color transformation engine. The result, we believe, is the foundation for a scalable, efficient, distributed RIP system that can meet current and future RIP requirements for a wide range of commercial digital presses.

Automating server tracking for data centers

Asset management in state of the art data centers is still a manual process. An automated system to track the location of servers and create a real-time inventory would not only improve operations management but also reduce operational costs. This demo shows a novel application of location aware wireless networks, RFID technology and visualization software integrated in end-to-end system to track servers in a data center.