Hassan Alam

Multiple Classifier Combination for Character Recognition: Revisiting the Majority Voting System and Its Variations

In recent years, strategies based on combination of multiple classifiers have created great interest in the character recognition research community. A huge number of complex and sophisticated decision combination strategies have been explored by researchers. However, it has been realized recently that the comparatively simple Majority Voting System and its variations can achieve very robust and often comparable, if not better, performance than many of these complex systems. In this paper, a review of various Majority Voting Systems and their variations are discussed, and a comparative study of some of these methods is presented for a typical character recognition task.

Structured and unstructured document summarization:design of a commercial summarizer using Lexical chains

The process of summarizing documents is becomingincreasingly important in the light of recent advances indocument creation/distribution technology, and theresulting influx of large numbers of documents in everyday life. This paper presents a document summarizer thatcombines document analysis, structural decomposition,XML representation and lexical chain analysis. Theproposed summarizer is compared to three commerciallyavailable summarizers and it is shown that it produceseither comparable or better summaries overall.

Automatic summarization of Web content to smaller display devices

Web documents usually have complicated layouts and the overall information content can be huge. All these documents are designed for viewing in large screen devices, such as a computer monitor. In recent times, a large number of small screen portable devices, such as personal digital assistants (PDA) and cellular phones, have been made available for mobile browsing. Viewing a Web page originally written for large screen devices using these very small screen devices can be extremely cumbersome. This paper discusses this issue of small viewing form factor of electronics devices from the perspective of Web browsing and proposes an approach to automatically summarize and transform Web documents into a meaningful, readable and above all, browsable format.

Web page summarization for handheld devices: a natural language approach

Summarization of web pages is a very interesting topicfrom both academic and commercial point of view.Academically, it is challenging to create a summary of adocument (e.g. a web page) that is highly structured andhas multi-media components in it. From the commercialpoint of view, it is advantageous to summarize web pagesto be viewed in small display devices such as PDAs andcell phones. Summarization not only makes web browsingand navigation easier, but it makes browsing faster ascomplete web pages need not be downloaded beforeviewing. In this paper, a novel combination of naturallanguage and non-natural language based summarizationtechniques have been used to automatically generate anintelligent re-authored display of web pages in real time.

FaxAssist : An automatic routing of unconstrained fax to email location

This paper describes a method for automatically routing unconstrained faxes to mail recipients. Incoming faxes usually have many formats, variation in name representation, and can be either machine or hand printed. Our approach synthesized the use of multiple OCR engines, the geometry based name location, and error correcting name matching. Candidates for a name match are determined by a number of components such as the qualification of the name image, the confidence of the name text, and the confidence of the keyword that appears in front of the name. Multiple name matches are then filtered out based on the weight toward the location of the name in the document and the distance between the keyword and the name, etc. As a result, our project accurately routes faxes to the right destination or sends the fax to an operator if the recipient cannot be recognized with sufficient confidence. An initial prototype fax routing system has been deployed and tested at DARPA. In this paper we discuss our approach, the results of testing at a live site and directions for the future.

A pair-wise decision fusion framework: recognition of human faces

Automatic detection and recognition offaces is now becoming an important tool in securiQ and suweillance. One of the most promising techniques for recognizing faces is the use of Support Vector Machines (SVMs). However, SVMs are essentially tw-class classifiers or dichotomizers, and applying them IO recognize multiple classes of faces requires post- recogni:ion decision fusion to arrive at the final overall decision. In this paper, we propose such a decision fusion framewrk, and show chat such a pair-nise classification framework can be totally generic irrespective of the chosen classifier. Application o/ this framework to the face recognition problem has resulted in encouraging perl/ormance.

Extending a broad-coverage parser for a general NLP toolkit

With the rapid growth of real world applications for NLP systems, there is a genuine demand for a general toolkit from which programmers with no linguistic knowledge can build specific NLP systems. Such a toolkit should have a parser that is general enough to be used across domains, and yet accurate enough for each specific application. In this paper, we describe a parser that extends a broad-coverage parser, Minipar (Lin, 2001), with an adaptable shallow parser so as to achieve both generality and accuracy in handling domain specific NL problems. We test this parser on our corpus and the results show that the accuracy is significantly higher than a system that uses Minipar alone.

Second Guessing a Commercial’Black Box’ Classifier by an’In House’ Classifier: Serial Classifier Combination in a Speech Recognition Application

We describe how an ’in house’ classifier can enhance the performance of a commercial ’black box’ classifier using the classic serial multiple classifier combination scheme. It is now acknowledged by the classifier combination community that parallel or hybrid decision fusion algorithms, in general, outperform serial combination schemes. However, classifier combination using techniques that use class labeling, ranking or probability estimators need access to low level information supplied by all of the participating classifiers. Unfortunately, in many commercial applications the classifier is often a’black box’, which implies that it is not possible to manipulate the low level information regarding classification for these classifiers. In many such cases, a serial classifier combination model provides the only practical method to improve classification. In this paper, we present such an application in speech recognition.

Exploring a hybrid of support vector machines (SVMs) and a heuristic based system in classifying web pages

Due to the proliferation of various types of devices used to browse the web and the shift of document access via web interfaces, it is now becoming very important to classify web pages into pre-selected types. This often forms the pre-processing stage of a number of web applications. However, classification of web pages is known to be a difficult problem because it is inherently difficult to identify specific features of web pages that are distinct and therefore it is equally difficult to use a set of heuristics to accomplish this. This paper describes a solution to the problem by combining a heuristic based system and a Support Vector Machine (SVM). It is found that such a hybrid system is able to perform at a very high accuracy when compared to using SVMs on their own.

A Pattern Recognition approach to automated XBRL extraction

Using example-based Pattern Recognition methods and combining years of developing both EDGAR filings and natural language processing software, BCL Technologies has developed SmartXBRL©, a simplified and automated way to create a compliant XBRL document. In this paper we describe the methods adopted to identify and extract the face financial tables, Document and Entity Information (DEI), Parenthetical, and financial Notes from a 10-Q financial document.