Untitled Document

.: Seminars :.

August 10th, 2005

Pierrick Legrand, CICESE-INRIA
1. Non-linear Wavelet Coefficients Pumping.
2. Denoising by Genetic Algorithms and Fractal Analysis.

October 25th, 2005 (13:00hrs "Optics Department")

Bir Bhanu, UC-Riverside
1. Synthesis of Recognition Systems.

Designing object detection and recognition systems that work effectively in the real world is a challenging task due to various factors including the high complexity of the systems, the dynamically changing environment and factors such as occlusion, clutter, articulation, and various noises that make the extraction of reliable features quite difficult. Furthermore, features useful in the detection and recognition of one kind of object or in the processing of one kind of imagery may not be effective in the detection and recognition of another kind of object or in the processing of another kind of imagery. Thus, the detection and recognition system needs thorough overhaul when applied to other types of images different from the one for which the system was designed. This is very uneconomical and requires highly trained experts. This talk will present evolutionary computational techniques to automate the synthesis and analysis of object detection and recognition systems. With learning incorporated, the resulting recognition systems will be able to automatically discover features according to the type of objects and images to which they are applied. The ultimate goal of the learning approaches is to lower the cost of designing object detection and recognition systems and build robust and flexible systems that can be adapted to varying task requirements. Examples from various domains will be presented.

November 30th, 2005 (17:00hrs "Electronics Department")

Humberto Sossa, CIC (Centro de Investigación en Computación, IPN)
1. Memorias asociativas: Avances recientes y aplicaciones.

Resumen: Una memoria asociativa es un dispositivo útil para asociar patrones. Cada par de patrones x y y, (x,y) se llama asociación. Dadas p
asociaciones, una memoria asociativa M es construida con estas p asociaciones. A este conjunto de asociaciones se le llama conjunto
fundamental de asociaciones (CFA). En esta plática se describe un nuevo modelo de memoria asociativa basado en la operación del conocido operador
mediana. Esto permite, por un lado, la construcción y uso de la memoria a través de un sólo operador externo. Evita también el uso de los llamados
kerneles para atacar el problema del ruido mezclado. Se dan condiciones formales de recuperación perfecta del CFA, sin ruido y ruido en los patrones
de entrada x. Se muestran también algunos ejemplos de aplicación de las nuevas herramientas.