International Journal of Pattern Recognition and Artificial Intelligence

When Pattern Recognition is aimed at human task automatization, such as the interpretation of visual data, it leads to Knowledge-Based Pattern Recognition (KBPR) systems. The design of a system, which contains a mixture of human knowledge and numerical data treatments, implies some tasks like Knowledge Acquisition (KA), dividing a problem into sub-problems or identifying different perceptual functions. Such problems cannot be solved by pure PR methods. Improvement of performances in KBPR systems is achieved through a better understanding of the cognitive processes which are involved during data interpretation and introspection which provides thinking aloud data. This paper will focus on problems concerned with knowledge as they are encountered in the design of expert systems for visual data interpretation. KA is considered as a necessary task for the system designer. Useful information is collected from the expert, however KA cannot provide a complete description of the human recognition process, directly usable for automatic numerical data interpretation. Verbal descriptions are lacking when introspection cannot reveal cognitive states such as those, for example, where perceptual adaptation is activated. The acquired knowledge consists of verbal descriptions as well as non-verbal knowledge provided by the analysis of the expert in action. The combination of acquired human knowledge and numerical data treatments is illustrated through examples where low level interpretation is achieved on task-adapted data representations. The need to design intermediate tools for making the dialogue more efficient or to define operational data representations is emphasized.

This paper surveys a long term project, aimed at providing a general methodology for building up and maintaining an expert system oriented to Pattern Recognition problems. The methodology makes use of an integrated set of modules, performing different functions but sharing a common knowledge representation scheme. In particular, a learning module allows to acquire the knowledge automatically from a set of examples and another module performs sophisticated reasoning, on the basis of the available knowledge, during the recognition phase.

The acoustic-phonetic decoding of sentences is a major bottleneck in continuous speech recognition. Our group has been working in this area for the past twelve years. In order to improve the accuracy of present phonetic decoders, we are now developing an expert system called APHODEX that should emulate the competence of an experienced spectrogram reader. After a brief description of the methodology used to elucidate the expert's knowledge and strategies, we focus on the expert's approach and results. Finally we explain how this expertise has been implemented; some results from the present experimental version of our system are given.

This paper presents an application of rule-based expert systems to thermal and structural analysis, part of mechanical engineering design, in electronic system design. The purpose of these expert systems is to serve as a consultant or design reviewer. These design expert systems are capable of performing both symbolic reasoning and numeric computation. The advantages of expert systems over conventional systems such as user friendliness, explanation and help facilities, and easy development and maintenance are also shown in this paper. Using system-subsystem tree structure several interrelated expert systems can be integrated into one system, and it is flexible for future expansion or modification. Object-oriented approach is also discussed, which with its unique representation of object could be a useful tool for engineering design.

Expert systems and image understanding have traditionally been considered as two separate application fields of artificial intelligence (AI). In this paper it is shown, however, that the idea of building an expert system for image understanding may be fruitful. Although this paper may serve as a framework for situating existing works on knowledge-based vision, it is not a review paper. The interested reader will therefore be referred to some recommended survey papers in the literature.

In this paper the use of digital pattern recognition in ultrasonic nondestructive evaluation (NDE) of materials is examined. Emphasis is placed on identifying effective features from time and frequency domains, correlation functions and impulse responses to classify aluminium plate specimen into three major defect geometries: flat and angular cuts, and circular holes. A multi-stage classification procedure is developed which can further determine the angles and sizes for defect characterization and classification. The research clearly demonstrates that modern pattern analysis can significantly improve the nondestructive material evaluation capability of the ultrasonic methods.

This paper addresses problems in underwater seismic image analysis and interpretation. We discuss basic issues involving acquisition and processing of signals due to reflections and multiple reflections from the sea bed and the underlying media. Aspects of compensating for the dynamic heave component present in the signals, using conventional and parallel Kalman filtering are discussed. Considerations for modeling the heave process are treated briefly. A number of information extraction algorithms that are central to the interpretation process are covered. Emphasis is given to delay and amplitude parameter estimation using a combined cross-relation-minimum variance filter, linearized recursive estimation and an event enhancement filter. Ingredients of a knowledge-based interpretation system for image analysis and interpretation are discussed.

This paper starts with a brief discussion of so-called wavelet transforms, i.e., decompositions of arbitrary signals into localized contributions labelled by a scale parameter. The main features of the method are first illustrated through simple mathematical examples. Then we present the first applications of the method to the recognition and visualisation of characteristic features of speech and of musical sounds.