Michael A. Zmuda

Automatic generation of morphological sequences

Morphological sequences (algorithms or programs) are generated using an evolutionary approach. A population of morphological sequences is manipulated and expanded in discrete steps. At each time-step two tasks are initiated--program discovery and program construction. The discovery phase searches for short morphological sequences which extract novel features. Program composition utilizes these sequences, which are partial solutions, to form increasingly effective sequences. The composition phase selects pairs of sequences and combines them into extended sequences which capture spatial relationships. The enhanced population serves as the basis for another phase of discovery and composition. Several demonstrations illustrate the systems ability to synthesize and integrate feature extraction routines.

Adaptive search for morphological feature detectors

A closed-loop hybrid learning system that facilitates the automatic design of a multi-class pattern recognition system is described. The design process has three phases: feature detector generation feature set selection and classification. In the first phase a large population of feature detectors based on morphological erosion and hit-or-miss operators is generated randomly. From this population an optimized subset of features is selected using a novel application of genetic algorithms. The selected features are then used to initialize a generalized Hamming neural network that performs image classification. This network provides the means for self-organizing the set of training patterns into additional subclasses this in turn dynamically alters the number of detectors and the size of the neural network. The design process uses system errors to gradually refine the set of feature vectors used in the classification subsystem. We describe an experiment in which the hybrid learning paradigm successfully generates a machine that distinguishes ten classes of handprinted numerical characters.

Multi-resolution feature extraction from Gabor filtered images

In this paper, we describe a hybrid learning system which combines a genetic algorithm with a neural network to classify grayscale images. The system operates on multi-resolution images which are formed by applying Gabor filters to a set of input images. The genetic algorithm evolves morphological probes that sample the multi-resolution images, and the perceptron algorithm then evaluates the extracted features. We demonstrate the use of this system by discriminating images of model tanks from other military vehicles. A multiplicity of accurate solutions, consisting of sparse morphological probes, are generated.<<ETX>>

Approaches to synthesizing image processing programs

Machine learning techniques are examined as a means of automatically generating image processing programs. Nonstructured techniques such as discovery systems and evolutionary processes are studied because they facilitate the exploration of enormous search spaces without a detailed knowledge base. The success of these methods depends on the algorithm representation and the effectiveness of performance evaluation. Mathematical morphology provides an algebraic representation which is powerful and challenging to program. The qualitative aspects of effective performance measures are also discussed.<<ETX>>

Efficient software techniques for morphological image processing with desktop computers

The basis of a system for processing binary images with the operations of mathematical morphology is described. This system exploits the properties of mathematical morphology to minimize computing time and storage requirements. Images are stored in data structures which are memory-efficient and allow several images to be processed simultaneously. Techniques are also presented for efficiently storing globally sized structuring elements. These ternary images are stored in data structures which utilize an adaptive window to provide storage for a 2M X 2N specification space in an optimal M X N data structure. This representation provides efficient storage, retrieval, and comparison of generalized structuring elements.

Applications of learning strategies to pattern recognition

Experiments are described with a hybrid learning system that automates the generation of multiclass pattern recognition systems. The learning system utilizes genetic algorithms to formulate a small set of feature detectors and an adaptive neural network to classify feature vectors. The experiments utilize a training set of handprinted characters and a pool of randomly generated morphological hit-or-miss detectors. A major problem is selecting a small subset of cooperating detectors from a large, easily generated pool of detectors. A new method of selecting detectors from a pool is presented that utilizes a modified version of crossover operators from genetic algorithms. This new crossover approach for generating feature detectors is evaluated by comparing it with random selection and with a restricted random mutation approach. The system adaptively adjusts the size of detector sets and the corresponding number of neural net nodes.© (1991) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Feature extraction using morphological analysis of multiresolution gray-scale images

In this paper, we discuss an initial effort to generate pattern recognizers using a multi- resolution Gabor stack of filtered images and a simple evolutionary search algorithm. The generated feature detectors are sets of pixel detectors that measure intensities and pass these values as feature vectors to neural net classifiers. We demonstrate the use of random search to solve a discrimination problem in which tank images are separated from other military vehicle images. The techniques and results used in this paper for discrimination of grey-scale images are reminiscent of similar approaches used to generate pattern recognizers for binary images. A sparse sampling of the Gabor image stack, using only 35 pixel detectors, produces feature vectors which are readily separated by linear perceptrons.

Biological Evolution as a Paradigm for Performance Driven Design Processes

The application of evolutionary processes to the problem of automated design is explored. A feature detector component for a pattern recognition system is used as an example of the automated design process. Results obtained from applying the resultant detector system to the problem of character recognition are discussed.

Generating feature detectors with discovery algorithms

Traditional techniques for extracting features from images are highly structured processes which require human experts to convert their intuition and experience into algorithms that solve problems such as image classification, target recognition, or assembly line inspection. Intelligent systems such as rule-based expert systems have been used to assist in the development process; however, these approaches still require significant human intervention to achieve acceptable results. This paper describes a system called MORPH which synthesizes complex feature extraction routines using only classification information provided by the image analyst. This system generates a multiplicity of very accurate solutions for several classification tasks.<<ETX>>

Computational complexity of synthesizing target recognition programs

An important research objective is to develop systems which automatically generate target recognition programs. This paper presents evidence that such general goals are not feasible. Specifically, the problem of automatically synthesizing target recognition programs is shown to be NP-Complete. The intractability of this problem motivates a problem specification which is tolerant of errors. Although easier, this too is shown to be NP-Complete. These results indicate that automatic target recognition has computational limitations which are inherent in the problem specification, and not necessarily a lack of clever system designs.