Roberto A. Hexsel

Particle filtering in the Hough space for instrument tracking

In this paper we present a real-time tracking system of surgical instruments in laparoscopic operations. We combine Condensation tracking, with the Hough Transform in order to obtain an efficient and accurate tracking. The Condensation algorithm performs well in heavy clutter, and the Hough Transform is robust under illumination changes, occlusion and distractions. The Hough array is computed using the gradient direction image obtained by means of a Principal Component Analysis. This improves accuracy in the determination of edge orientation and speeds up computation of the Hough Transform. The experiments on image sequences of actual laparoscopic surgical operations show that the instrument tip is located even in the presence of smoke, occlusions or motion blurring.

IRIS RECOGNITION USING ADABOOST AND LEVENSHTEIN DISTANCES

This paper presents an efficient IrisCode classifier, built from phase features which uses AdaBoost for the selection of Gabor wavelets bandwidths. The final iris classifier consists of a weighted contribution of weak classifiers. As weak classifiers we use three-split decision trees that identify a candidate based on the Levenshtein distance between phase vectors of the respective iris images. Our experiments show that the Levenshtein distance has better discrimination in comparing IrisCodes than the Hamming distance. Our process also differs from existing methods because the wavelengths of the Gabor filters used, and their final weights in the decision function, are chosen from the robust final classifier, instead of being fixed and/or limited by the programmer, thus yielding higher iris recognition rates. A pyramidal strategy for cascading filters with increasing complexity makes the system suitable for real-time operation. We have designed a processor array to accelerate the computation of the Levenshtein distance. The processing elements are simple basic cells, interconnected by relatively short paths, which makes it suitable for a VLSI implementation.

The performance of pollution control victim cache for embedded systems

In this paper we explore the design space of data caches looking for the combinations of design parameters that produce the best results at the smallest sizes. We introduce a technique named Pollution Control Victim Cache (PCVC) which improves the Pollution Control Cache (PCC), is simpler and performs better. Our simulations were run on the SimpleScalar suite running the Commbench benchmarks. Our results indicate that Victim Caches and Stream Buffers do not perform well in the small systems we simulated. The PCC and our PCVC have shown promising results.

Approximate String Matching for Iris Recognition by Means of Boosted Gabor Wavelets

This paper presents an efficient IrisCode classifier, built from phase features which uses AdaBoost for the selection of Gabor wavelets bandwidths. The final iris classifier consists of a weighted contribution of weak classifiers. As weak classifiers we use 3-split decision trees that identify a candidate based on the Levenshtein distance between phase vectors of the respective iris images. Our experiments show that the Levenshtein distance has better discrimination in comparing IrisCodes than the Hamming distance. Our process also differs from existing methods because the wavelengths of the Gabor filters used, and their final weights in the decision function, are chosen from the robust final classifier, instead of being fixed and/or limited by the programmer, thus yielding higher iris recognition rates. A pyramidal strategy for cascading filters with increasing complexity makes the system suitable for realtime operation.

Translating full duplexity into capacity gains for the high-priority traffic classes of IEEE 802.11

In this paper we present PbP-EDCA, a Full Duplex (FD) MAC protocol that operates according to the high-priority access rules of the standard IEEE 802.11 EDCA. With PbP-EDCA, nodes that win a contention perform a sequence of narrow-band orthogonal transmissions rather than one single wide-band. PbP-EDCA employs the Carrier Sense Multiple Access/Collision Avoidance (CSMA/CA) to avoid sequences starting at the same time and takes advantage of self-interference free radios to enable multiple simultaneous orthogonal transmissions to/from a single-radio access point. By exploiting FD and the EDCA contention values standardized for high-priority traffic classes, we show that PbP-EDCA can scale the gains claimed by single-link FD proposals in a network-wide domain. Our results indicate one scenario in which PbP-EDCA can achieve about 3× the capacity of the IEEE 802.11 half duplex MAC.

A minimalist cache coherent MPSoC designed for FPGAs

We describe the design and VHDL implementation of a cache coherent MPSoC named minimalist cache coherent MPSoC (MCCM). The system comprises one to eight MIPS-I processors, coherent primary data caches, memory management units, memory controller and the interconnection. We present a detailed account of the implementation, focusing on the shared memory subsystem. A simple benchmark is used to assess the overall system functionality. We compared the size of our design to that of a LEON3-based multiprocessor and found that a four-core LEON3 system needs roughly the same amount of logic/state as a six to eight cores MCCM.