Showing papers in "Artificial Intelligence Review in 1998"
TL;DR: Different models of genetic operators and some mechanisms available for studying the behaviour of this type of genetic algorithms are revised and compared.
Abstract: Genetic algorithms play a significant role, as search techniques for handling complex spaces, in many fields such as artificial intelligence, engineering, robotic, etc. Genetic algorithms are based on the underlying genetic process in biological organisms and on the natural evolution principles of populations. These algorithms process a population of chromosomes, which represent search space solutions, with three operations: selection, crossover and mutation.
Under its initial formulation, the search space solutions are coded using the binary alphabet. However, the good properties related with these algorithms do not stem from the use of this alphabet; other coding types have been considered for the representation issue, such as real coding, which would seem particularly natural when tackling optimization problems of parameters with variables in continuous domains. In this paper we review the features of real-coded genetic algorithms. Different models of genetic operators and some mechanisms available for studying the behaviour of this type of genetic algorithms are revised and compared.
1,190 citations
TL;DR: A new pattern recognition system to classify large numbers of plankton images detected in real time by the Video Plankton Recorder, a towed underwater video microscope system, achieving 95% classification accuracy on six plankton taxa taken from nearly 2,000 images.
Abstract: Plankton form the base of the food chain in the ocean and are fundamental to marine ecosystem dynamics. The rapid mapping of plankton abundance together with taxonomic and size composition is very important for ocean environmental research, but difficult or impossible to accomplish using traditional techniques. In this paper, we present a new pattern recognition system to classify large numbers of plankton images detected in real time by the Video Plankton Recorder (VPR), a towed underwater video microscope system. The difficulty of such classification is compounded because: 1) underwater images are typically very noisy, 2) many plankton objects are in partial occlusion, 3) the objects are deformable and 4) images are projection variant, i.e., the images are video records of three-dimensional objects in arbitrary positions and orientations. Our approach combines traditional invariant moment features and Fourier boundary descriptors with gray-scale morphological granulometries to form a feature vector capturing both shape and texture information of plankton images. With an improved learning vector quantization network classifier, we achieve 95% classification accuracy on six plankton taxa taken from nearly 2,000 images. This result is comparable with what a trained biologist can achieve by using conventional manual techniques, making possible for the first time a fully automated, at sea-approach to real-time mapping of plankton populations.
113 citations
TL;DR: Experimental results using the improved end-effector showed that the fruits were harvested successfully with no damage, and the functions to detect the fruit position and the air pressure in the pad were installed, so that theruits were harvested regardless of the length of their peduncle.
Abstract: Two types of robotic end-effectors capable of harvesting tomato fruits were manufactured based on the physical properties of tomato plant and tested The first prototype end-effector consisted of two parallel plate fingers and a suction pad The fingers pick a fruit off at the joint of its peduncle after the suction cup singulates it by vacuum from other fruits in the same cluster From the results of harvesting experiment, the end-effector could not harvest fruits with a short peduncle because the fruits were detached from the suction pad before they were gripped by the fingers Therefore, the second prototype in which the functions to detect the fruit position and the air pressure in the pad were installed, was made, so that the fruits were harvested regardless of the length of their peduncle Experimental results using the improved end-effector showed that the fruits were harvested successfully with no damage
104 citations
TL;DR: A blood spot detection neural network was trained, tested, and evaluated entirely on eggs with blood spots and grade A eggs, and these accuracy levels were sufficient to produce graded samples that would exceed the USDA requirements.
Abstract: A blood spot detection neural network was trained, tested, and evaluated entirely on eggs with blood spots and grade A eggs. The neural network could accurately distinguish between grade A eggs and blood spot eggs. However, when eggs with other defects were included in the sample, the accuracy of the neural network was reduced. The accuracy was also reduced when evaluating eggs from other poultry houses. To minimize these sensitivities, eggs with cracks and dirt stains were included in the training data as examples of eggs without blood spots. The training data also combined eggs from different sources. Similar inaccuracies were observed in neural networks for crack detection and dirt stain detection. New neural networks were developed for these defects using the method applied for the blood spot neural network development.
The neural network model for blood spot detection had an average accuracy of 92.8%. The neural network model for dirt stained eggs had an average accuracy of 85.0%. The average accuracy of the crack detection neural network was 87.8%. These accuracy levels were sufficient to produce graded samples that would exceed the USDA requirements.
69 citations
TL;DR: Some considerations and examples of robotics development for plant production are presented followed by a description of the key components of plant production robots.
Abstract: Applying robotics in plant production requires the integration of robot capabilities, plant culture, and the work environment. Commercial plant production requires certain cultural practices to be performed on the plants under certain environmental conditions. Some of the environmental conditions are mostly natural and some are modified or controlled. In many cases, the required cultural practices dictate the layout and materials flow of the production system. Both the cultural and environmental factors significantly affect when, where and how the plants are manipulated. Several cultural practices are commonly known in the plant production industry. The ones which have been the subject of robotics research include division and transfer of plant materials in micropropagation, transplanting of seedlings, sticking of cuttings, grafting, pruning, and harvesting of fruit and vegetables. The plants are expected to change their shape and size during growth and development. Robotics technology includes many sub-topics including the manipulator mechanism and its control, end-effector design, sensing techniques, mobility, and workcell development. The robots which are to be used for performing plant cultural tasks must recognize and understand the physical properties of each unique object and must be able to work under various environmental conditions in fields or controlled environments. This article will present some considerations and examples of robotics development for plant production followed by a description of the key components of plant production robots. A case study on developing a harvesting robot for an up-side-down single truss tomato production system will also be described.
56 citations
TL;DR: The paper presents an overview of fuzzy logic modeling techniques, its applications to biological and agricultural systems and an example showing the steps of constructing a fuzzy logic model.
Abstract: Fuzzy logic is a powerful concept for handling non-linear, time-varying, adaptive systems. It permits the use of linguistic values of variables and imprecise relationships for modeling system behavior. The paper presents an overview of fuzzy logic modeling techniques, its applications to biological and agricultural systems and an example showing the steps of constructing a fuzzy logic model.
55 citations
TL;DR: It is argued that CBR, psychology and legal theory have complementary contributions to understanding similarity, and what each offers is described, to establish criteria for assessing existing CBR systems in law.
Abstract: Case-based Reasoning (CBR) began as a theory of human cognition, but has attracted relatively little direct experimental or theoretical investigation in psychology. However, psychologists have developed a range of instance-based theories of cognition and have extensively studied how similarity to past cases can guide categorization of new cases. This paper considers the relation between CBR and psychological research, focussing on similarity in human and artificial case-based reasoning in law. We argue that CBR, psychology and legal theory have complementary contributions to understanding similarity, and describe what each offers. This allows us to establish criteria for assessing existing CBR systems in law and to establish what we consider to be the crucial goals for further research on similarity, both from a psychological and a CBR perspective.
37 citations
TL;DR: An image analysis procedure was developed to quantify morphological characteristics of convolutions in individual cotton fibers without pre-tensioning or orientation requirements, and results agreed with visual inspection in 89.6% of the tested images.
Abstract: An image analysis procedure was developed to quantify morphological characteristics of convolutions in individual cotton fibers without pre-tensioning or orientation requirements. The image of each fiber was captured by a PC-based color imaging system using a conventional microscope. Ends of individual cotton fibers were glued on a microscope slide without any tension or straightening. A modified watershed technique was implemented to identify individual convolution segments, which were defined as sections of the fiber bordered by two neighboring convolutions. Length, area and perimeter of each convolution segment were measured directly from the image. Average width, shape factor and number of convolution segments in mm were calculated from the measured parameters. Performance of the image analysis algorithm was compared with visual inspection for number and position of convolution segments in three different varieties of cotton. Image analysis results agreed with visual inspection in 89.6% of the tested images.
20 citations
TL;DR: The development of a model that can be employed in the development of an intelligent tutoring system that is capable of offering remedial tutoring according to principles of remediation, a formalisation ofmedial interventions with intelligent tutoringsystems is developed.
Abstract: For successful teaching to take place an intelligent tutoring system has to be able to cope with any student errors that may occur during a tutoring interaction. Remedial tutoring is increasingly viewed as a central part of the overall tutoring process, and recent research calls for adaptive remedial tutoring. This paper discusses the issues of remedial tutoring that have been proposed or implemented to support efficient remedial tutoring. These issues serve to uncover any underlying principles of remediation that govern remedial tutoring with intelligent tutoring systems. In order to incorporate these principles of remediation into intelligent tutoring systems development processes this paper continues with the development of a model that can be employed in the development of an intelligent tutoring system that is capable of offering remedial tutoring according to these principles. This model is a formalisation of remedial interventions with intelligent tutoring systems. To demonstrate how the model can be employed in developing an intelligent tutoring system, INTUITION, the implementation of an existing business simulation game, has been developed. This paper concludes with an illustration of how the model for remedial operations provides for remedial tutoring within INTUITION. The evaluation of INTUITION shows that the model for remedial operations is a useful method for providing efficient remedial tutoring.
20 citations
TL;DR: The validity of the image reconstruction algorithm was evaluated by processing several layered digital images of known shape and size and differences between the original and reconstructed images were 2–5% in terms of object size and 1–8% in Terms of shape.
Abstract: Confocal laser scanning microscopy (CLSM) is a noninvasive technique for evaluating the microstructure of foods and other materials. CLSM provides several sequential subsurface layers of two-dimensional (2-D) images. An image processing algorithm was developed to reconstruct these 2-D layers into a three-dimensional (3-D) network. Microstructure of fat globules in cheese was used as an example application. The validity of the image reconstruction algorithm was evaluated by processing several layered digital images of known shape and size. Differences between the original and reconstructed images were 2–5% in terms of object size and 1–8% in terms of shape.
17 citations
TL;DR: Practical generation of identification keys for biological taxa using a multilayer perceptron neural network is described, and it is demonstrated that such an Artificial Neural Network Key (ANNKEY) can identify more than half (52.9%) of the species in this genus, even though data for one character is completely missing.
Abstract: In this paper, practical generation of identification keys for biological taxa using a multilayer perceptron neural network is described. Unlike conventional expert systems, this method does not require an expert for key generation, but is merely based on recordings of observed character states. Like a human taxonomist, its judgement is based on experience, and it is therefore capable of generalized identification of taxa. An initial study involving identification of three species of Iris with greater than 90% confidence is presented here. In addition, the horticulturally significant genus Lithops (Aizoaceae/Mesembryanthemaceae), popular with enthusiasts of succulent plants, is used as a more practical example, because of the difficulty of generation of a conventional key to species, and the existence of a relatively recent monograph. It is demonstrated that such an Artificial Neural Network Key (ANNKEY) can identify more than half (52.9%) of the species in this genus, after training with representative data, even though data for one character is completely missing.
TL;DR: A novel system for grading oranges into three quality bands, according to their surface characteristics, is described, which contains state-of-the-art processing hardware, a novel mechanical design, and three separate algorithmic components.
Abstract: We describe a novel system for grading oranges into three quality bands, according to their surface characteristics. The system is designed to process fruit with a wide range of size (55–100 mm), shape (spherical to highly eccentric), surface coloration and defect markings. This application requires both high throughput (5–10 oranges per second) and complex pattern recognition. The grading is achieved by simultaneously imaging each item of fruit from six orthogonal directions as it is propelled through an inspection chamber. In order to achieve the required throughput, the system contains state-of-the-art processing hardware, a novel mechanical design, and three separate algorithmic components. One of the key improvements in this system is a method for recognising the point of stem attachment (the calyx) so that it can be distinguished from defects. A neural network classifier on rotation invariant transformations (Zernike moments) is used to recognise the radial colour variation that is shown to be a reliable signature of the stem region. The succession of oranges processed by the machine constitute a pipeline, so time saved in the processing of defect free oranges is used to provide additional time for other oranges. Initial results are presented from a performance analysis of this system.
TL;DR: A multiagent system for studying in vitro cell motion and results, coming from an existing prototype, show different types of cell behavior during cell migration, based on cell nuclei analysis.
Abstract: This paper presents a multiagent system for studying in vitro cell motion. A typical application on the wound closure process is presented to illustrate the possibilities of the system, where different image sequences will be treated. The motion issue involves three aspects: image segmentation, object tracking and motion analysis. The current system version focuses mainly on the image segmentation aspect. A general agent model has been designed, which will be further expanded to include tracking and motion analysis behaviors as well. The agents integrate three basic behaviors: perception, interaction and reproduction. The perception evaluates pixels upon static and motion-based criteria. The interaction behavior allows two agents to merge or to negotiate parts of regions. The negotiation can be seen as a segmentation refinement process done by the agents. Finally, the reproduction behavior defines an exploration strategy of the images. Agents can start other agents around them, or they can duplicate themselves in the next frame. The frames are processed in pipeline, where previous information is used to treat the current frame. One unique agent model exist. Agents are specialized on execution time according to their goals. The results, coming from an existing prototype, show different types of cell behavior during cell migration, based on cell nuclei analysis.
TL;DR: A review of some recently developed techniques in the field of natural language processing, including a discussion about the underlying similarities between some of these systems and mention two approaches to the evaluation of statistical language processing systems.
Abstract: We present a review of some recently developed techniques in the field of natural language processing. This area has witnessed a confluence of approaches which are inspired by theories from linguistics and those which are inspired by theories from information theory: statistical language models are becoming more linguistically sophisticated and the models of language used by linguists are incorporating stochastic techniques to help resolve ambiguities. We include a discussion about the underlying similarities between some of these systems and mention two approaches to the evaluation of statistical language processing systems.
TL;DR: A computer-vision based system was used and was successful in recognizing individual shreds even when shreds were touching or overlapping, and shred lengths calculated from the processed images compared very well with those measured manually.
Abstract: A computer-vision based system was used to obtain images of shredded cheese. The images were processed by morphological transformation algorithms such as dilation and erosion to smooth the image edge contours. The smoothed image was skeletonized. Cheese shred lengths were determined from skeletonized images using syntactic methods. This method was successful in recognizing individual shreds even when shreds were touching or overlapping. Shred lengths calculated from the processed images compared very well with those measured manually.
TL;DR: A realistic view is taken of the prospects for medium term progress and some observations are made concerning the direction this might profitably take in terms of principal connectionist learning methods and neurological modeling trends.
Abstract: This paper is a survey of some recent connectionist approaches to the design and development of behaviour-based mobile robots. The research is analysed in terms of principal connectionist learning methods and neurological modeling trends. Possible advantages over conventionally programmed methods are considered and the connectionist achievements to date are assessed. A realistic view is taken of the prospects for medium term progress and some observations are made concerning the direction this might profitably take.
TL;DR: Results show that the NUFZY model with the fast OLS training can perform quite well in predicting both lettuce growth and greenhouse temperature, and is able to incorporate the human knowledge in this approach, and to deduce any interpretable rules that describe the systems' behavior.
Abstract: A hybrid neuro-fuzzy approach called the NUFZY system, which embeds fuzzy reasoning into a triple-layered network structure, has been developed to identify nonlinear systems. A set of membership functions at the input layer is partially linked with a layer of rules, using pre-set parameters. By means of a simplified centroid of gravity defuzzification method, the output becomes linear in the weights. Therefore, very fast estimation of the weight parameters can be achieved by using the orthogonal least squares (OLS) method, which also provides a method to efficiently remove the redundant fuzzy rules from the prototype rule base of the NUFZY system. In this paper, the NUFZY system is applied to identify lettuce growth and greenhouse temperature from real experimental data.
Results show that the NUFZY model with the fast OLS training can perform quite well in predicting both lettuce growth and greenhouse temperature. In contrast to the mechanistic modeling procedures, the neuro-fuzzy approach offers an easier route and a fast way to build the nonlinear mapping of inputs and outputs. In addition, the resulting internal network structure of the NUFZY system is a self-explanatory representation of fuzzy rules. Under this frame, it is a perspective that one is able to incorporate the human knowledge in this approach, and, hopefully, to deduce any interpretable rules that describe the systems‘ behavior.
TL;DR: A review of the application of the quad tesseral representation to support spatial reasoning is presented, finding that it linearises multi-dimensional space, while still providing for the description of individual objects within that space and the relationships that may exist between those objects.
Abstract: A review of the application of the quad tesseral representation to support spatial reasoning is presented. The principal feature of the representation is that it linearises multi-dimensional space, while still providing for the description of individual objects within that space and the relationships that may exist between those objects (in any direction and through any number of dimensions). In addition the representation is supported by an arithmetic which allows the manipulation (translation etc.) of spatial objects. Consequently, when incorporated into a spatial reasoning system, all necessary processing can be implemented as if in only one dimension. This offers two significant advantages over more conventional multi-directional approaches to spatial reasoning. Firstly, many of the concerns associated with the exponential increase in the number or relations that need to be considered (as the number of dimensions under consideration increases) are no longer relevant. Secondly, the computational cost of manipulating and comparing spatial objects remains static at its one dimensional level, regardless of the number of dimensions under consideration.
TL;DR: The approach involves reasoning about the reuse of both qualitative and mathematical models for physiological processes, and constructs quantitative simulation models for the postharvest behaviour of agricultural products.
Abstract: In this paper, we present an approach to automated modelling of physiological processes occurring during postharvest distribution of agricultural products. The approach involves reasoning about the reuse of both qualitative and mathematical models for physiological processes, and constructs quantitative simulation models for the postharvest behaviour of agricultural products. The qualitative models are used to bridge the gap between the modeller‘s knowledge about the physiological phenomenon and the mathematical models. The qualitative models are represented by knowledge graphs, that are conceptual graphs containing only causal relations, aggregation relations, and generalisation relations between domain quantities. The relationships between the mathematical models and the qualitative models are explicitly represented in application frames.
The automated modelling task consists of two subtasks. In the first subtask, Qualitative Process Analysis, a process structure graph is constructed using the qualitative models as building blocks. The process structure graph is a qualitative description of the phenomenon under study, that contains the processes that are responsible for the behaviour of the phenomenon. The process structure graph serves as a focus for the second subtask, Simulation Model Construction. This subtask uses a library of mathematical models to compose a quantitative simulation model that corresponds to the process structure graph constructed in the first subtask. The approach is illustrated with the construction of a model for the occurrence of chilling injury in bell peppers.
TL;DR: This paper aims at providing a smooth introduction to autoepistemic logic, stressing its motivation and basic concepts, and shows a more understandable, operational method for determining expansions.
Abstract: The subject of nonmonotonic reasoning is reasoning with incomplete information. One of the main approaches is autoepistemic logic in which reasoning is based on introspection. This paper aims at providing a smooth introduction to this logic, stressing its motivation and basic concepts. The meaning (semantics) of autoepistemic logic is given in terms of so-called expansions which are usually defined as solutions of a fixed-point equation. The present paper shows a more understandable, operational method for determining expansions. By improving applicability of the basic concepts to concrete examples, we hope to make a contribution to a wider usage of autoepistemic logic in practical applications.