Jerneja Gros

Evolution of the Information-Retrieval System for Blind and Visually-Impaired People

Blind and visually-impaired people face many problems in interacting with information retrieval systems. State-of-the-art spoken language technology offers potential to overcome many of them. In the mid-nineties our research group decided to develop an information retrieval system suitable for Slovene-speaking blind and visually-impaired people. A voice-driven text-to-speech dialogue system was developed for reading Slovenian texts obtained from the Electronic Information System of the Association of Slovenian Blind and Visually Impaired Persons Societies. The evolution of the system is presented. The early version of the system was designed to deal explicitly with the Electronic Information System where the available text corpora are stored in a plain text file format without any, or with just some, basic non-standard tagging. Further improvements to the system became possible with the decision to transfer the available corpora to the new web portal, exclusively dedicated to blind and visually-impaired users. The text files were reformatted into common HTML/XML pages, which comply with the basic recommendations set by the Web Access Initiative. In the latest version of the system all the modules of the early version are being integrated into the user interface, which has some basic web-browsing functionalities and a text-to-speech screen-reader function controlled by the mouse as well.

Spoken Language Resources at LUKS of the University of Ljubljana

This paper presents the Slovene-language spoken resources that were acquired at the Laboratory of Artificial Perception, Systems and Cybernetics (LUKS) at the Faculty of Electrical Engineering, University of Ljubljana over the past ten years. The resources consist of:• isolated-spoken-word corpora designed for phonetic research of the Slovene spoken language;• read-speech corpora from dialogues relating to air flight information;• isolated-word corpora, designed for studying the Slovene spoken diphthongs;• Slovene diphone corpora used for text-to-speech synthesis systems;• a weather forecast speech database, as an attempt to capture radio and television broadcast news in the Slovene language; and• read- and spontaneous-speech corpora used to study the effects of the psycho physical conditions of the speakers on their speech characteristics.All the resources are accompanied by relevant text transcriptions, lexicons and various segmentation labels. The read-speech corpora relating to the air flight information domain also are annotated prosodically and semantically. The words in the orthographic transcription were automatically tagged for their lemma and morphosyntactic description. Many of the mentioned speech resources are freely available for basic research purposes in speech technology and linguistics. In this paper we describe all the resources in more detail and give a brief description of their use in the spoken language technology products developed at LUKS.

Principal directions of synthetic exact filters for robust real-time eye localization

The alignment of the facial region with a predefined canonical form is one of the most crucial steps in a face recognition system. Most of the existing alignment techniques rely on the position of the eyes and, hence, require an efficient and reliable eye localization procedure. In this paper we propose a novel technique for this purpose, which exploits a new class of correlation filters called Prinicpal directions of Synthetic Exact Filters (PSEFs). The proposed filters represent a generalization of the recently proposed Average of Synthetic Exact Filters (ASEFs) and exhibit desirable properties, such as relatively short training times, computational simplicity, high localization rates and real time capabilities. We present the theory of PSEF filter construction, elaborate on their characteristics and finally develop an efficient procedure for eye localization using several PSEF filters. We demonstrate the effectiveness of the proposed class of correlation filters for the task of eye localization on facial images from the FERET database and show that for the tested task they outperform the established Haar cascade object detector as well as the ASEF correlation filters.

Homer II-man-machine interface to internet for blind and visually impaired people

HOMER II is a voice-driven text-to-speech system developed for blind or visually impaired persons for reading Slovenian texts. Users can obtain texts from the Internet site of the Association of Slovenian Blind and Visually Impaired Persons Societies from their Electronic Information System where they can find daily newspapers, some novels and other information. The system consists of four main modules. The first module enables Internet communication, retrieves text to a local disc and converts it to a standard form. The input interface manages the keyboard entry and/or speaker independent speech recognition. The output interface performs speech synthesis of a given text and in addition prints the same text magnified to the screen. The user dialog is responsible for the user friendly communication and controls other tasks of the system. Homer II was ported from Linux to the MS Windows 9x/ME/NT/2000 operating systems. For the best performance it uses multi-threading and other advantages of the 32-bit environment. Further versions of the HOMER system with even more advanced dialogue modules and some basic World Wide Web browsing functionality will represent an important tool in the distance learning and teaching process for the impaired persons using academic networks.

The Phonetic SMS Reader

The Phonectic SMS Reader is a speech technology test application that enables sending SMS messages and receiving them in voice format. The application can be viewed as a new value-added service for a mobile telephony operator. The test system described includes two mobile telephones, one for intercepting SMS messages and another one for calling the message receivers. The text message is converted to speech by a text-to-speech system.

HOMER: a voice-driven text-to-speech system for the blind

HOMER is a voice-driven text-to-speech system developed for blind or visually impaired persons for reading the Slovenian texts. Users can obtain texts from the special corpora organised on the computer network server at the information centre of the Association of the Slovenian Blind and Visually Impaired Persons. The system consists of three main modules. The text-to-speech module enables speech synthesis from an arbitrary Slovenian text input, the speech recognition module performs speaker independent isolated word recognition and the dialogue module controls the different tasks of the HOMER system and obtains texts from the source text corpora. Presently, the system runs under Linux and requires a Pentinum/133 PC with minimum 32 MB of RAM and an additional standard 16 bit sound card.

Composite score normalization for face verification

Similarity scores, which form the basis for identity inference in biometric verification systems, typically exhibit statistical variations. These variations are caused by so-called miss-matched conditions, in which the enrollment and probe samples were acquired, and are common to most application domains of biometric verification systems ranging from forensics to smart-home environments. To mitigate these variations, score normalization techniques are usually used. Examples of these techniques include the z-norm, the t-norm or the zt-norm. In this paper we study two-step normalization techniques, such as the zt-norm, and propose a new way of implementing such techniques. Specifically, we propose to implement the first step of the two-step procedure off-line in a non-parametric manner, while the second step is kept unchanged and, hence, performed parametrically. As shown in our face verification experiments, the proposed composite scheme can improve upon the performance of parametric normalization techniques, without an increase in computational complexity, as this is the case with pure non-parametric normalization techniques.

FST-Based Pronunciation Lexicon Compression for Speech Engines

Finite-state transducers are frequently used for pronunciation lexicon representations in speech engines, in which memory and processing resources are scarce. This paper proposes two possibilities for further reducing the memory footprint of finite-state transducers representing pronunciation lexicons. First, different alignments of grapheme and allophone transcriptions are studied and a reduction in the number of states of up to 30% is reported. Second, a combination of grapheme-to-allophone rules with a finite-state transducer is proposed, which yields a 65% smaller finite-state transducer than conventional approaches.

Fusion of discriminative and generative scoring criteria in GMM-based speaker verification

The aim of this paper is to demonstrate the complementarity of different scoring methods used in speaker verification. To show that, we implemented two different scoring methods on top of the joint factor analysis model. The results on the telephone part of the NISTs SRE 2008 core condition show that significant increase in performance can be achieved by fusing likelihood ratioand support vector machine-based scores.

Language Model Representations for the GOPOLIS Database

The formation of a domain-oriented sentence corpus by sentence pattern rules is described. The same rules were transformed into word networks to serve as a language model within a HTK based speech recognition system. The performance of the word network language model was compared to the one of the bigram model.