Visual Recognition System Research Articles

A validity measure for fuzzy clustering

The authors present a fuzzy validity criterion based on a validity function which identifies compact and separate fuzzy c-partitions without assumptions as to the number of substructures inherent in the data. This function depends on the data set, geometric distance measure, distance between cluster centroids and more importantly on the fuzzy partition generated by any fuzzy algorithm used. The function is mathematically justified via its relationship to a well-defined hard clustering validity function, the separation index for which the condition of uniqueness has already been established. The performance of this validity function compares favorably to that of several others. The application of this validity function to color image segmentation in a computer color vision system for recognition of IC wafer defects which are otherwise impossible to detect using gray-scale image processing is discussed.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Cooking Works Automation by the Cooking Robot. 1st Report. Visual Recognition Expert System of Cooking Robot Using Neural Network.

This paper presents a visual recognition expert system for cooking robots. The "cooking robot" is considered to be a robot which chooses ingredients and tools and cooks automatically, once ordered what to cook. For the development of this kind of cooking robot in the near future, we believe the function of automatic recognition of the cooking state and the ingredient will become important. Regarding this point, this paper proposes a visual recognition system which can evaluate the doneness of meat and kinds of meat. The system recognizes the meat using image-processed data and pattern-matching data given by the neural network and synthesizes these data under the rules written in the expert system. Experimental results show that this system can distinguish the doneness of meat and the kind of meat excellently.

Space and time bounds on indexing 3D models from 2D images

Model-based visual recognition systems often match groups of image features to groups of model features to form initial hypotheses, which are then verified. In order to accelerate recognition considerably, the model groups can be arranged in an index space (hashed) offline such that feasible matches are found by indexing into this space. For the case of 2D images and 3D models consisting of point features, bounds on the space required for indexing and on the speedup that such indexing can achieve are demonstrated. It is proved that, even in the absence of image error, each model must be represented by a 2D surface in the index space. This places an unexpected lower bound on the space required to implement indexing and proves that no quantity is invariant for all projections of a model into the image. Theoretical bounds on the speedup achieved by indexing in the presence of image error are also determined, and an implementation of indexing for measuring this speedup empirically is presented. It is found that indexing can produce only a minimal speedup on its own. However, when accompanied by a grouping operation, indexing can provide significant speedups that grow exponentially with the number of features in the groups. >

Knowledge based interpretation of images: a biomedical perspective

We formulate Dempster Shafer Belief functions in terms of Propositional Logic, using the implicit notion of provability underlying Dempster Shafer Theory. Given a set of propositional clauses, assigning weights to certain propositional literals enables the Belief functions to be explicitly computed using Network Reliability techniques. Also, the logical procedure corresponding to updating Belief functions using Dempster's Rule of Combination is shown. This analysis formalizes the implementation of Belief functions within an Assumption-based Truth Maintenance System (ATMS). We describe the extension of an ATMS-based visual recognition system, VICTORS, with this logical formulation of Dempster Shafer theory. Without Dempster Shafer theory, VICTORS computes all possible visual interpretations (i.e. all logical models) without determining the best interpretation(s). Incorporating Dempster Shafer theory enables optimal visual interpretations to be computed and a logical semantics to be maintained.

Automatic compilation of parallelism in visual object recognition

The paper presents the results of investigations into techniques for automatically generating concurrent implementations of visual object recognition systems from high-level visual domain descriptions. Visual knowledge can be represented using ‘schemata’-object-centred recognition models that intentionally represent classes of related objects. Schemata form hierarchies based on relationships such as composition and function. Schema-based representations contain implicit concurrencies that can be exploited by a recognition system called a ‘recognition network’. The paper presents the design of a simple schema representation language, and shows how it can be automatically transformed into concurrent code. This is done by building a compiler that takes schemata as input and produces a recognition network, in occam, as output. The design of the compiler involves several novel features. The existence of large transputer arrays makes it possible to execute the resulting code with high physical, as well as conceptual, parallelism.

A parallel algorithm for determining two-dimensional object positions using incomplete information about their boundaries

Extraction of two-dimensional object locations using current techniques is a computationally intensive process. In this paper a parallel algorithm is presented that can specify the location of objects from edge streaks produced by an edge operator. Best-first searches are carried out in a number of non-interacting and localized edge streak spaces. The outcome of each search is a hypothesis. Each edge streak votes for a single hypothesis; it may also take part in the formation of other hypotheses. A poll of the votes determined the stronger hypotheses. The algorithm can be used as a front end to a visual pattern recognition system where features are extracted from the hypothesized object boundary or from the area localized by the hypothesized boundary. Experimental results from a biomedical domain are presented.

Deficits in human visual spatial attention following thalamic lesions.

There has been speculation concerning the role that thalamic nuclei play in directing attention to locations in visual space [Crick, F. (1984) Proc. Natl. Acad. Sci. USA 81, 4586-4590]. We measured covert shifts of visual attention in three patients with unilateral thalamic hemorrhages shortly after the lesion and after a 6-month recovery period. The experiment measured reaction time to targets that occurred at locations to which attention had been cued (valid trials) or at a currently unattended location (invalid trials). Although the patients showed no deficits in visual fields with perimetry and no neglect in the 6-month follow-up, we found slow reaction times for targets on the side contralateral to the lesion whether or not attention had been cued to that location. Deficits have also been found in this task with cortical and midbrain lesions, but the patterns of performance are quite different. The results with thalamic patients suggest they have a specific deficit in the ability to use attention to improve the efficiency of processing visual targets contralateral to the lesion (engage operation). This finding is in accord with hypotheses of a thalamic link between cortical visual attention and pattern recognition systems proposed by Crick.

A Multi-Microprocessor-Based Control System for Assembly of Electronic Components

Abstract Design experiences of a project on building a multi-microprocessor-based automation control system for the assembly of electronic components is reported in this paper. The control system consists of four subsystems, which are the supervisory controller, the visual pattern recognition system, the display unit and the servo system. The emphasis was put on the design of a fast, cost-effective visual information-processing unit. For this, a new type of frame grabber together with noise-insensitive image processing algorithms are incorporated by employing the TMS 4161 dual ported memory. To verify the proposed design concepts, an experimental control system is developed and applied for the semiconductor die-bonding machine.

Concepts in vertebrate neuroethology

Neuroethology is concerned with the analysis of the neural substrates and mechanisms that underlie behaviour. In 1951, Tinbergen implied the following goals of Ethophysiology, today known as Neuroethology: (1) recognition and localization of behaviourally meaningful stimuli, e.g. key stimuli; (2) sensorimotor interfacing and feedback interactions; (3) modulations according to internal states and acquired information; (4) motor pattern generation; and (5) ontogenetic and phylogenetic aspects. Using the toad's ( Bufo bufo) visually-guided prey-catching behaviour as an example, experimental strategies and concepts of vertebrate Neuroethology can be demonstrated: (i) visual space is mapped in the brain in a multiple manner; (ii) specification of neurons results from inhibitory and/or excitatory interactions among and within space-mapped neuronal networks; (iii) specialized neurons, e.g. those with stimulus recognition and localization properties, show information processing that takes place in functional units of cell assemblies (so-called ‘neuronal machines’ or ‘brain chips’); (iv) activation of motor pattern generators for different behavioural actions may require coincidence of inputs from different combinations of specialized neurons that serve as command elements and together form a command system; (v) a command system can be regarded as a sensorimotor interface fulfilling tasks with respect to (a) visual pattern recognition and localization, (b) command functions, by initiating the activation of the motor pattern generation system, (c) motor pattern generation through participation in the temporal sequence of the positive feedback from the motor system, and (d) integration of modulatory inputs according to internal states; (vi) the motor pattern generation circuitry consists of a neuronal network capable of producing a consistent distribution of excitation and inhibition and whose output has privileged access to the required motoneuronal pools; (vii) the basic principles of prey selection and motor pattern generation of prey capture emerge in the post-metamorphic juvenile toad with the transition to land and without previous prey experience. During early ontogeny the acuity of sensory discrimination and the performance of motor patterns are subject to maturation paralleled by neural differentiations in the bbrain. performance of motor patterns are subject to maturation paralleled by neural differentiations in the brain. those are programmed and, in amphibians, presumably less dependent on the interaction with the environment than in mammals. This of course does not exclude the possibility of modulating, modifying, specifying and/or extending the visual pattern recognition system of toads due to internal states and individual experience.

Developmental dyslexia: a cognitive experimental analysis of phonological, morphemic, and visual impairments

Abstract The results of a cognitive experimental analysis of the reading functions of four developmentally dyslexic subjects are presented. Data are considered on an individual basis in an attempt to specify the primary source of disturbance in each case, and the manner in which the normal course of reading development has been distorted. Subject RO showed an effect on the analytic functioning of the visual processor, which affected speed of non-word reading, but appeared not to have prevented the construction of a normal visual word recognition system. In subject LT a phonological impairment was implicated, and this was associated with a pattern of performance similar to that of adult phonological dyslexics. Subjects GS and SE gave evidence of serial letter-by-letter reading (impaired wholistic function of the visual processor), which produced a pattern not unlike that of adult surface dyslexics (effects of spelling regularity, regularisation errors). The manner in which these impairments might disrupt t...

The evoked cortical potential in macular degeneration.

ABSTRACT: Macular degeneration causes a profound loss of visual acuity due to inadequate blood circulation in the capillaries of the choroid plexus and pigment epithelium of the important pattern‐recognizing area of the retina. This disease is the cause of half the blindness in people over the age of 65 and of nearly one‐third of all blindness. In this study an important new technique was added to six tests already in use to assess retinal disease. Data on 10 patients were collected by the following methods: 1) optic disc photography; 2) fluorescein angiography; 3) visual acuity testing; 4) visual fields; 5) foveal sensitivity; 6) electroretinography with automatic analysis of the record by use of an electronic sampling filter and a specially built computer to measure the total electrochemical energy released by the retina; and 7) the evoked cortical response (as indicated in the electroencephalogram) to the stimulus of an alternating chess‐board pattern. This new technique tests the total system for visual pattern recognition. The results indicate that the evoked cortical response is the best objective test of visual acuity.