Costas S. Xydeas

Cross-band pixel selection in multiresolution image fusion

The work described in this paper focuses on cross band pixel selection as applied to pixel level multi-resolution image fusion. In addition, multi-resolution analysis and synthesis is realized via QMF sub-band decomposition techniques. Thus cross-band pixel selection is considered with the aim of reducing the contrast and structural distortion image artifacts produced by existing wavelet based, pixel level, image fusion schemes. Preliminary subjective image fusion results demonstrate clearly the advantage which the proposed cross-band selection technique offers, when compared to conventional area based pixel selection.

The Manchester Multimedia Information System

A Multimedia Information System (MMIS) is a repository for all types of electronically representable data (O’Docherty et al., 1990). Conventional databases provide a large set of operations for retrieval of simple data types. The simplest way of extending this to multimedia objects is to store and retrieve on the basis of a few manually entered associated attributes or links. The true potential of multimedia databases is realised when a rich set of operations is provided to allow transparent manipulation of data objects of all media. This can best be achieved through content retrieval, based on the automatic interpretation of medium objects. Automatic content retrieval avoids the problems of inconsistency, subjectivity and the labour-intensiveness of manual entry. MMISs with content retrieval will have wide application in industry, medicine, education and the military. The Multimedia Group at Manchester University have prototyped a MMIS for the content retrieval of images in a specific application domain and are developing a second system that includes content retrieval of text and documents. Hereafter, the term instance is used to refer to medium objects that are intended for interpretation in our system.

Detecting the skew angle in document images

Abstract In the field of document image analysis, accurate detection and removal of intrinsic skew is of paramount importance as a first step in the processing of document images. Here we present an efficient scheme for detecting the degree of misalignment in a document page. The proposed algorithm operates directly on the raw digitised image and is shown to achieve a high skew detection accuracy for mixed mode document formats which contain typewritten text, cursive script, line-art and photographic pictures. We also discuss efficiency considerations for a practical real-time hardware implementation of the algorithm. Furthermore, in a practical document image processing environment, it is necessary to process documents that are landscape or portrait oriented. In this context we present an algorithm which determines the page orientation prior to skew detection.

On improving vector excitation coders through the use of spherical lattice codebooks (SLCs)

Two novel techniques for use in VXC (vector excitation coding) speech coders are presented. The first enables massive excitation codebooks (>or=20 b) to be used at realizable complexities by using a novel spherical lattice codebook for the excitation codebook. The second technique is a generalization of the gain-optimized error measure which allows any number of gains to be calculated for each excitation vector. This multiple-gain VXC can be thought of as a hybrid between multipulse and VXC.<<ETX>>

Low bit rate coding using an interpolated zinc excitation model

Prototype excitation vocoding has been of particular interest, particularly at bit rates below 4 kbits/sec where conventional analysis by synthesis (AbS) LPC based coders fail to deliver the speech quality required in a number of new voice communication applications and services. The present authors propose a new time domain, interpolated zinc function prototype excitation technique (IZFPE) for communications quality speech at about 2.0 kb/s. In this method a prototype pitch segment is selected every 20 ms. Each prototype segment is then modelled by a zinc orthogonal function which is derived by employing a closed loop AbS procedure (ZFE-AbS). Successive prototype zinc excitations are then interpolated such that the characteristics of the resulting excitation signal evolve slowly with time.<<ETX>>

A long history quantization approach to scalar and vector quantization of LSP coefficients

LHQ (long history quantization), when applied in conjunction with scalar quantization for the coding of LSP (line spectrum pair) coefficients in a CELP (code excited linear prediction) system, offers transparent quantization at an average bit rate of 25 bits/frame. Certain improvements to the LHQ-scalar quantization scheme are presented which further reduce the average LSP bit rate to about 22 bits/frame. In addition, a novel LHQ-vector quantization scheme is proposed, which allows the transparent quantization of LPC coefficients with only 19 bits/analysis frame.<<ETX>>

Robust speech recognition using fuzzy matrix quantisation, neural networks and Hidden Markov models

In this paper a new approach to robust speech recognition using Fuzzy Matrix Quantisation, Hidden Markov Models and Neural Networks is presented and tested when speech is corrupted by car noise. Thus two new robust isolated word speech recognition (IWSR) systems called FMQ/HMM and FMQ/MLP, are proposed and designed optimally for operation in a variety of input SNR conditions. The schemes and associated system training methodologies result into a particularly high recognition performance at input SNR levels as low as 5 and 0 dBs.

Conversion of scanned documents to the open document architecture

The paper presents a system for the conversion of scanned documents into the open document architecture. Unlike previous work in this field the authors use a combination of evidence sources to achieve greater robustness to document defects and noise introduced in the scanning process. Furthermore, they use optical character recognition in conjunction with other forms of image analysis as a means of detecting document structure. This enables enhanced document feature extraction and improved performance. They demonstrate the performance of the system on a specific class of input document.<<ETX>>

Hierarchical classification method and its application in shape representation

In this paper we describe a technique for performing shaped-based content retrieval of images from a large database. In order to be able to formulate such user-generated queries about visual objects, we have developed an hierarchical classification technique. This hierarchical classification technique enables similarity matching between objects, with the position in the hierarchy signifying the level of generality to be used in the query. The classification technique is unsupervised, robust, and general; it can be applied to any suitable parameter set. To establish the potential of this classifier for aiding visual querying, we have applied it to the classification of the 2-D outlines of leaves.

A comparison of system architectures for intelligent document understanding

Abstract Intelligent document understanding (IDU) is the process of converting scanned document images into a high level representation which describes the documents layout and logical structure, in addition to providing its information content. In this paper we discuss IDU in general and address a specific problem within this domain concerning the extraction of the layout structure of pages from a technical journal. Three different architectural approaches to accomplishing this task are proposed. Firstly we describe a novel document understanding system (System A) which exploits a hybrid bottom-up/top-down control architecture. The system uses a variety of image processing algorithms in a bottom-up manner. Conversely, a system based on a pure top-down architecture (System B) is then proposed which produces a segmentation of the page via projection profile analysis and achieves classification of image regions via procedural deduction. Finally, an alternative top-down architecture (System C) is described in which an optimised segmentation scheme is applied to produce partitioned blocks. These are then classified in a goal driven manner using a decision tree. A comparison of the three systems is made by measuring system performance on images obtained from a specific class of input document. The performance of document understanding systems has been quantified in terms of an object identification rate and the percentage of column area successfully interpreted. Using these measures, System A has given superior results to the two top-down systems presented. System A also performs significantly better than a previously reported top-down system operating on a comparable problem (Viswanathan, 1990).