A futuristic account of the ICDAR'07 Ninthconference is presented. It attracted 2740 participants, including 2480 teleattendants. the program committee selected 185 papers from 640 submissions. thirty sessions we...
详细信息
ISBN:
(纸本)0818678984;0818678992
A futuristic account of the ICDAR'07 Ninthconference is presented. It attracted 2740 participants, including 2480 teleattendants. the program committee selected 185 papers from 640 submissions. thirty sessions were grouped according to nine major themes: text recognition, graphics, devices, datamining and warehousing, sublimation, oral documents, documents, epistechnology, and history of optical character recognition.
this paper presents a prototypical digital library service. It integrates machinelearning tools and techniques in order to make effective, efficient and economically feasible the process of capturing the information ...
详细信息
this paper presents a prototypical digital library service. It integrates machinelearning tools and techniques in order to make effective, efficient and economically feasible the process of capturing the information that should be stored and indexed by content in the digital library. In fact, information capture is one of the main bottleneck when building a digital library, since it involves complex patternrecognition problems, such as document analysis, classification and understanding. Experimental results show that learning systems can solve effectively and efficiently all these problems.
Assisted document recognition systems have to integrate automatic recognition, manual edition and incremental learning in a single interactive environment. this paper raises the question of the organization of these t...
详细信息
Assisted document recognition systems have to integrate automatic recognition, manual edition and incremental learning in a single interactive environment. this paper raises the question of the organization of these three kinds of operations. When an analyzer has the ability to improve with use, there is a tradeoff between the benefits of enhancing the accuracy of automatic analysis, and the additional time spent in interacting for feedback communication. the global cost depends then on the sequence of processed entities, and on the relevance of the learning transactions. Notations are introduced to describe the evolution of a recognition session, and possible organization strategies are discussed. then a cost model is presented to allow the comparison between different organization schemes. We describe some concrete experiments of cost measures withthe ApOFIS font identification tool and the ScanWorX OCR;the first results show that a user-driven approach can potentially save substantial effort in the recognition process, in comparison withmachine-driven systems.
Almost alt document analysis approaches need to perform a global analysis of the page orientation as a separate process at an early stage. ft would be preferable to estimate the orientation locally after page segmenta...
详细信息
ISBN:
(纸本)0818678984;0818678992
Almost alt document analysis approaches need to perform a global analysis of the page orientation as a separate process at an early stage. ft would be preferable to estimate the orientation locally after page segmentation and classification, when more knowledge about the different regions is available, In this paper, a novel local skew estimation method is presented that takes advantage of the information available after flexible and efficient page segmentation and classification methods have been applied to the document image. the proposed method accurately estimates the orientation of individual text regions by efficiently analysing the arrangement of background space contained in them. No assumption is made of the existence Of a uniform or dominant orientation in the document. the whole process is very efficient, as only the regions of text are considered and the points used for the angle estimation are already available as by-products of previous document analysis stages.
the proceedings contains 78 papers from the 1997 IEEE internationalconference on Tools with Artificial Intelligence. Topics discussed include: neural networks;knowledge representation and reasoning;artificial intelli...
详细信息
the proceedings contains 78 papers from the 1997 IEEE internationalconference on Tools with Artificial Intelligence. Topics discussed include: neural networks;knowledge representation and reasoning;artificial intelligence;software engineering;genetic algorithms;logic based reasoning systems;natural language processing;vision and patternrecognition;optimization problem solving tools;evolutionary computation;object-oriented methodologies;intelligent agents;knowledge based systems;intelligent user interfaces;datamining;and machinelearning.
the architecture of a system for reading machine-printed documents in known predefined tabular-data layout styles is described. In these tables, textual data are presented in record lines made up of fixed-width fields...
详细信息
the architecture of a system for reading machine-printed documents in known predefined tabular-data layout styles is described. In these tables, textual data are presented in record lines made up of fixed-width fields. the system performs these steps: copes with multiple tables per page;identifies records within tables;segments records into fields;and recognizes characters within fields, constrained by field-specific contextual knowledge. Obstacles to good performance on tables include small print, tight line-spacing, poor-quality text, and line-art or background patterns that touch the text.
this work proposes an original approach to musical score recognition, a particular case of high-level document analysis. In order to overcome the limitations of existing systems, we propose an architecture which allow...
详细信息
this work proposes an original approach to musical score recognition, a particular case of high-level document analysis. In order to overcome the limitations of existing systems, we propose an architecture which allows for a continuous and bidirectional interaction between high-level knowledge and low-level data, and which is able to improve itself over time by learning. this architecture is made of three cooperating layers, one made of parameterized feature detectors, another working as an object-oriented knowledge repository and the other as a supervising Bayesian metaprocessor. Although the implementation is still in progress, we show how this architecture is adequate for modeling and processing knowledge.
datamining algorithms including machinelearning, statistical analysis, and patternrecognition techniques can greatly improve our understanding of data warehouses that are now becoming more widespread. In this paper...
详细信息
ISBN:
(纸本)0780337573
datamining algorithms including machinelearning, statistical analysis, and patternrecognition techniques can greatly improve our understanding of data warehouses that are now becoming more widespread. In this paper;we focus on classification algorithms and review the need for multiple classification algorithms. We describe a system called MLC++, which was designed to help choose the appropriate classification algorithm for a given dataset by making it easy to compare the utility of different algorithms on a specific dataset of interest. MLC++ not only provides a work-bench for such comparisons, but also provides a library of C++ classes to aid in the development of new algorithms, especially hybrid algorithms and multi-strategy algorithms. Such algorithms are generally hard to code from scratch. We discuss design issues, interfaces to other programs, and visualization of the resulting classifiers.
the authors have investigated a neural network (NN) as a tool of PD patternrecognitionthese years. this paper describes several aspects of NN through authors experiments. the NN system uses input signal of φ-q-n di...
详细信息
the authors have investigated a neural network (NN) as a tool of PD patternrecognitionthese years. this paper describes several aspects of NN through authors experiments. the NN system uses input signal of φ-q-n distribution patterns. the φ-q-n distribution consists of the pulse count [n] versus pulse height [q] and phase angle [φ]. the learning characteristics and recognition characteristics of the NN were investigated. the basic characteristics of recognition capability for combined pattern signal input was shown. the outline of the effectiveness of the NN system for partial discharge recognition and a new method of data standardization of phase-shifted φ-q-n data was proposed and its effectiveness for PD patternrecognition performance were described.
Practical tools for knowledge discovery from databases must be efficient enough to handle large data sets found in commercial environments. Attribute-oriented induction has proved to be a useful method for knowledge d...
详细信息
Practical tools for knowledge discovery from databases must be efficient enough to handle large data sets found in commercial environments. Attribute-oriented induction has proved to be a useful method for knowledge discovery. three algorithms are AOI, LCHR and GDBR. We have implemented efficient versions of each algorithm and empirically compared them on large commercial data sets. these tests show that GDBR is consistently faster than AOI and LCHR. GDBR's times increase linearly with increased input size, while times for AOI and LCHR increase non-linearly when memory is exceeded. through better memory management, however, AOI can be improved to provide some advantages.
暂无评论