Semi-supervised learning for relevance feedback on image retrieval tasks

Relevance feedback approaches have been established as an important tool for interactive search, enabling users to express their needs. However, in view of the growth of multimedia collections available, the user efforts required by these methods tend to increase as well, demanding approaches for reducing the need of user interactions. In this context, this paper proposes a semi-supervised learning algorithm for relevance feedback to be used in image retrieval tasks. The proposed semi-supervised algorithm aims at using both supervised and unsupervised approaches simultaneously. While a supervised step is performed using the information collected from the user feedback, an unsupervised step exploits the intrinsic dataset structure, which is represented in terms of ranked lists of images. Several experiments were conducted for different image retrieval tasks involving shape, color, and texture descriptors and different datasets. The proposed approach was also evaluated on multimodal retrieval tasks, considering visual and textual descriptors. Experimental results demonstrate the effectiveness of the proposed approach.

End-to-side loop neurorrhaphy: axonal comparative stereological study

Accidents or diseases can affect the peripheral part of the nervous system, which raises clinical and surgical therapies, among others. In this context, the technique of end-to-side neurorrhaphy is a treatment option, yet its modification loop needs some additional efficacy studies. The purpose of this study was to compare, among rats, stereological results (axons volume density) after end-to-side neurorrhaphy and after end-to-side loop neurorrhaphy. Thirty Wistar rats were used, divided into six groups (five animals per group), consisting of two control groups (for the fibular and tibial nerves), two study groups for the fibular nerve (one with an end-to-side neurorrhaphy, and the other with an end-to-side loop neurorrhaphy) and two study groups for the tibial nerve (with an endto- side neurorrhaphy and the other one with an end-to-side loop neurorrhaphy). After 180 days, all groups were sacrificed for axonal stereological analysis (volume density) in distal nerve stumps. There was significant maintenance of neuronal-axonal density in the distal stumps to neurorrhaphy (p< 0.005) compared with the normal stumps. The end-to-side loop neurorrhaphy is a therapeutic option as suture technique after complete nerve section, in order to restore most of the axonal functional integrity.

Comments on two-loop Kac-Moody algebras

It is shown that the two-loop Kac-Moody algebra is equivalent to a two-variable-loop algebra and a decoupled β-γ system. Similarly WZNW and CSW models having as algebraic structure the Kac-Moody algebra are equivalent to an infinity of versions of the corresponding ordinary models and decoupled abelian fields.

A new deformation of W-infinity and applications to the two-loop WZNW and conformal affine Toda models

We construct a centerless W-infinity type of algebra in terms of a generator of a centerless Virasoro algebra and an abelian spin 1 current. This algebra conventionally emerges in the study of pseudo-differential operators on a circle or alternatively within KP hierarchy with Watanabe's bracket. Construction used here is based on a spherical deformation of the algebra W ∞ of area preserving diffeomorphisms of a 2-manifold. We show that this deformation technique applies to the two-loop WZNW and conformal affine Toda models, establishing henceforth W ∞ invariance of these models.

Excited fermion contribution to Z(0) physics at one loop

We investigate the effects induced by excited leptons at the one-loop level in the observables measured on the Ζ peak at LEP. Using a general effective Lagrangian approach to describe the couplings of the excited leptons, we compute their contributions to both oblique parameters and Ζ partial widths. Our results show that the new effects are comparable to the present experimental sensitivity, but they do not lead to a significant improvement on the available constraints on the couplings and masses of these states.

Projeto de controladores robustos para acionamento de um motor de indução trifásico

Pós-graduação em Engenharia Elétrica - FEIS

Simulations of the Frequency Modulated Atomic Force Microscope (FM-AFM) nonlinear control system

The Frequency Modulated - Atomic Force Microscope (FM-AFM) is apowerful tool to perform surface investigation with true atomic resolution. The controlsystem of the FM-AFM must keep constant both the frequency and amplitude ofoscillation of the microcantilever during the scanning process of the sample. However,tip and sample interaction forces cause modulations in the microcantilever motion.A Phase-Locked Loop (PLL) is used as a demodulator and to generate feedback signalto the FM-AFM control system. The PLL performance is vital to the FM-AFMperformace since the image information is in the modulated microcantilever motion.Nevertheless, little attention is drawn to PLL performance in the FM-AFM literature.Here, the FM-AFM control system is simulated, comparing the performancefor di erent PLL designs.

Nonlinear state estimation and control for chaos suppression in MEMS resonator

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)