No abstract

Extracting regions that are noticeably different from their surroundings, so called salient regions, is a topic of considerable interest for image retrieval. There are many current techniques but it has been shown that SIFT and MSER regions are among the best. The SIFT methods have their basis in linear scale-space but less well known is that MSERs are based on a non-linear scale-space. We demonstrate the connection between MSERs and morphological scale-space. Using this connection, MSERs can be enhanced to form a saliency tree which we evaluate via its effectiveness at a standard image retrieval task. The tree out-performs scale-saliency methods. We also examine the robustness of the tree using another standard task in which patches are compared across images transformations such as illuminant change, perspective transformation and so on. The saliency tree is one of the best performing methods.1 To properly preserve scale-space causality [4], these DoG filters D(x, y, σ) = (G(x, y, kσ) − G(x, y, σ)) * I(x, y) should be formed from the discrete approximation to the Gaussian kernel [5] but, in practice, [3] uses a seven-point approximation to the kernel applied in a multi-resolution framework.

Dynamic process simulation is widely used in teaching controller design, as it allows foreseeing the performance of different control configurations and controller tunings. Currently, most college-level controller design exercises that are based on simulation consider deterministic perturbations (i.e. steps or ramps). In real life however, processes are more likely to face fluctuating, random disturbances, so the use of stochastic simulation in controller tuning exercises would provide students with an experience closer to their future professional practice than that provided by deterministic simulation. However, public institutions attempting to use dynamic, stochastic simulators in teaching, are hindered by the need of buying licenses of simulation packages or specialized programming languages (such as Matlab), as there aren´t any dynamic, stochastic simulators available as downloadable freeware. This paper shows a dynamic, stochastic simulator with a friendly interface of a distillation tower, developed as an Excel macro. This simulator has the advantage that it can be used at no cost to educational institutions since Excel is almost universally known and used by college faculties.