The Bhattacharyya distance functional invariance classifier design method (BDFICDM) reaches an excellent tradeoff between the accuracy performance of nearly 100% and the time performance as much as there is available computational resources for parallelism. A serial implementation of the method is applied to an image composed of two synthetic and two Bradatz textures preserving its characteristics of being highly suitable to a parallel implementation. This implementation characterizes the pattern by composing independent cells of magnitudes evaluated through the interaction of spatially distributed elementary piece of information (EPI) using the Bhatacharyya distance similarity measure. The localized representation, EPI is composed of shifted frames sampled by a sequential process over a direction in order to preserve the pattern topological information. The representational extension of the functional pattern representation is its reduction focusing only on those EPI best candidates for generating invariance locations and using the graph structures for their representation. The BDF sample is classified within a Bayesian approach by comparing it to only those reduced pattern states of invariance, decreasing abruptly the number of needed interactions and comparisons. The procedure comprises the multiple frame, resolution, hypothesis and class approaches and the image representation used as input in the training and classification processes at the frame decomposition instance, is reduced through the use of the KL transform.
Template regularization embeds the problem of class separability. In the machine vision perspective, this problem is critical when a textural classification procedure is applied to non-stationary pattern mosaic images. These applications often present low accuracy performance due to disturbance of the classifiers produced by exogenous or endogenous signal regularity perturbations. Natural scene imaging, where the images present certain degree of homogeneity in terms of texture element size or shape (primitives) shows a variety of behaviors, especially varying the preferential spatial directionality. The space-time image pattern characterization is only solved if classification procedures are designed considering the most robust tools within a parallel and hardware perspective. The results to be compared in this paper are obtained using a framework based on multi-resolution, frame and hypothesis approach. Two strategies for the bank of Gabor filters applications are considered: adaptive strategy using the KL transform and fix configuration strategy. The regularization under discussion is accomplished in the pyramid building system instance. The filterings are steering Gaussians controlled by free parameters which are adjusted in accordance with a feedback process driven by hints obtained from sequence of frames interaction functionals pos-processed in the training process and including classification of training set samples as examples. Besides these adjustments there is continuous input data sensitive adaptiveness. The experimental result assessments are focused on two basic issues: Bhattacharyya distance as pattern characterization feature and the combination of KL transform as feature selection and adaptive criterion with the regularization of the pattern Bhattacharyya distance functional (BDF) behavior, using the BDF state separability and symmetry as the main indicators of an optimum framework parameter configuration.
This paper shows the new approach results in analyzing and classifying test images focusing on the differences among the existing spatial frame sequence modelings obtained from each region candidate or class. The used tool combination applied to analyze the classify the mosaic images consists of a bank of Gabor filters for decomposing the image and Gaussian filters for building the multi-resolution image representation on the filter bank outputs, and two classifiers: a Bayesian and a low-resolution Bhattacharyya distance RCE neural network classifiers. The training set of textures consists of Brodatz and synthetic patterns.
A scheme for comparative performance analysis of the Bayesian and the Bhattacharyya distance RCE neural network classifiers is presented. The experiments are performed on synthetic and Brodatz textures. The introduction of the new classifier aims at obtaining a better performance in classifying non-stationary multi-texture images. The two classification schemes are assessed on their localized data representation regarding the ability of extracting non- stationary information from the image. Low-resolution data representation is used to reduce the instability produced with the search for a better trade-off between accuracy and spatial classification performances.
CONTROLAB is an environment which integrates intelligent systems and control algorithms aiming at applications in the area of robotics. Within CONTROLAB, two neural network architectures based on the backpropagation and the recursive models are proposed for the implementation of a robust speaker-independent word recognition system. The robustness of the system using the backpropagation network has been largely verified through use by children and adults in totally uncontrolled environments such as large public halls for the exhibition of new technology products. Experimental results with the recursive network show that it is able to overcome the backpropagation network major drawback, the frequent generation of false alarms. In addition, within CONTROLAB, the trajectory to be followed by a robot arm under self-tuning control is determined by a system which uses either VGRAPH or PFIELD algorithms to avoid obstacles detected by the computer vision system. The performance of the second algorithm is greatly improved when it is applied under the control of a rule-based system. An application in which a SCARA robot arm is commanded by voice to pick up a specific tool placed on a table among other tools and obstacles is currently running. This application is used to evaluate the performance of each sub-system within CONTROLAB.
Systems for high-precision control of the trajectory to be followed by a robot arm grip need to properly model the interaction among the robot arm joints and to cope with the high-speed and nonlinearities of the arm dynamics. To solve this problem the use of a hardware accelerator, which is able to explore parallelism within multivariable self-tuning control algorithms, is proposed. The accelerator works as part of an integrated system which incorporates facilities of computer vision and robot arm trajectory definition. The computer vision sub-system recognizes the position of an object selected to be picked by the robot arm and the trajectory definition sub-system uses a neural network to define the angular position of the joints along the trajectory to be followed by the arm.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.