Research bed for unit selection based text to speech synthesis

The paper describes a modular, unit selection based TTS framework, which can be used as a research bed for developing TTS in any new language, as well as studying the effect of changing any parameter during synthesis. Using this framework, TTS has been developed for Tamil. Synthesis database consists of 1027 phonetically rich prerecorded sentences. This framework has already been tested for Kannada. Our TTS synthesizes intelligible and acceptably natural speech, as supported by high mean opinion scores. The framework is further optimized to suit embedded applications like mobiles and PDAs. We compressed the synthesis speech database with standard speech compression algorithms used in commercial GSM phones and evaluated the quality of the resultant synthesized sentences. Even with a highly compressed database, the synthesized output is perceptually close to that with uncompressed database. Through experiments, we explored the ambiguities in human perception when listening to Tamil phones and syllables uttered in isolation,thus proposing to exploit the misperception to substitute for missing phone contexts in the database. Listening experiments have been conducted on sentences synthesized by deliberately replacing phones with their confused ones.

Research on laser welding of cast Ni-based superalloy K418 turbo disk and alloy steel 42CrMo shaft

Exploratory experiments of laser welding cast Ni-based superalloy K418 turbo disk and alloy steel 42CrMo shaft were conducted. Microstructure of the welded seam was characterized by optical microscopy (OM), scanning electron microscopy (SEM), X-ray diffraction (XRD), energy-dispersive spectrometer (EDS). Mechanical properties of the welded seam were evaluated by microhardness and tensile strength testing. The corresponding mechanisms were discussed in detail. Results showed that the laser-welded seam had non-equilibrium solidified microstructures consisting of FeCr0.29Ni0.16C0.06 austenite solid solution dendrites as the dominant and some fine and dispersed Ni3Al gamma' phase and Laves particles as well as little amount of MC short stick or particle-like carbides distributed in the interdendritic regions. The average microhardness of the welded seam was relatively uniform and lower than that of the base metal due to partial dissolution and suppression of the strengthening phase gamma' to some extent. About 88.5% tensile strength of the base metal was achieved in the welded joint because of a non-full penetration welding and the fracture mechanism was a mixture of ductility and brittleness. The existence of some Laves particles in the welded seam also facilitated the initiation and propagation of the microcracks and microvoids and hence, the detrimental effects of the tensile strength of the welded joint. The present results stimulate further investigation on this field. (c) 2006 Elsevier B.V. All rights reserved.

Atlas of abundance and distribution patterns of ichthyoplankton from the Northeast Pacific Ocean and Bering Sea ecosystems: based on research conducted by the Alaska Fisheries Science Center (1972–1996)

This regional atlas summarizes and illustrates the distribution and abundance patterns of fish eggs and larvae of 102 taxa within 34 families found in the Northeast Pacific Ocean including the Bering Sea, Gulf of Alaska, and U.S. west coast ecosystems. Data were collected over a 20+ year period (1972–1996) by the Recruitment Processes Program of the Alaska Fisheries Science Center (AFSC). Ichthyoplankton catch records used in this atlas were generated from 11,379 tows taken during 100 cruises. For each taxon, general life history data are briefly summarized from the literature. Published information on distribution patterns of eggs and larvae are reviewed for the study area. Data from AFSC ichthyoplankton collections were combined to produce an average spatial distribution for each taxon. These data were also used to estimate mean abundance and percent occurrence by year and month, and relative abundance by larval length and season. Abundance from each tow was measured as catch per 10 m2 surface area. A larval distribution and abundance map was produced with a geographic information system using ArcInfo software. For taxa with identifiable pelagic eggs, distribution maps showing presence or absence of eggs are presented. Presence or absence of adults in the study area is mapped based on recent literature and data from AFSC groundfish surveys. Distributional records for adults and early life history stages revealed several new range extensions. (PDF file contains 288 pages.)

Growth of the buccaneer anchovy Encrasicholina punctifer off Mozambique, based on samples collected in research surveys

Growth of Encrasicholina punctifer (Engraulididae) is estimated, based on samples obtained during five surveys on Sofala Bank (central Mozambique). The estimated values of the von Bertalanffy growth function, TL∞=12 cm and K=2.0 yearˉ¹ are compared with results of growth studies elsewhere in the Indo-Pacific region, based mainly on the index Φ=log(sub)10 K+2*log(sub)10L∞ assumed to be constant within species. This comparison is rendered difficult by the uncertain taxonomical status of stolephorid anchovies.

Research on continuous speech recognition with small vocabulary based on biomimetic pattern recognition

Based on biomimetic pattern recognition theory, we proposed a novel speaker-independent continuous speech keyword-spotting algorithm. Without endpoint detection and division, we can get the minimum distance curve between continuous speech samples and every keyword-training net through the dynamic searching to the feature-extracted continuous speech. Then we can count the number of the keywords by investigating the vale-value and the numbers of the vales in the curve. Experiments of small vocabulary continuous speech with various speaking rate have got good recognition results and proved the validity of the algorithm.

Continuous speech research based on HyperSausage Neuron

In this paper, we presents HyperSausage Neuron based on the High-Dimension Space(HDS), and proposes a new algorithm for speaker independent continuous digit speech recognition. At last, compared to HMM-based method, the recognition rate of HyperSausage Neuron method is higher than that of in HMM-based method.

Research of Face Location System Based on Human Vision Simulations

In this paper we present a robust face location system based on human vision simulations to automatically locate faces in color static images. Our method is divided into four stages. In the first stage we use a gauss low-pass filter to remove the fine information of images, which is useless in the initial stage of human vision. During the second and the third stages, our technique approximately detects the image regions, which may contain faces. During the fourth stage, the existence of faces in the selected regions is verified. Having combined the advantages of Bottom-Up Feature Based Methods and Appearance-Based Methods, our algorithm performs well in various images, including those with highly complex backgrounds.

Continuous speech research based on HyperSausage Neuron

In this paper, we presents HyperSausage Neuron based on the High-Dimension Space(HDS), and proposes a new algorithm for speaker independent continuous digit speech recognition. At last, compared to HMM-based method, the recognition rate of HyperSausage Neuron method is higher than that of in HMM-based method.