Kent Ridge Digital Labs

Agency for Science, Technology and Research

National University of Singapore

Haizhou Li

Jiankang Wu

Shuanhu Bai

Jun Xu

Cuntai Guan

Ye Wang

Qibin Sun

Arcot Desai Narasimhalu

Changsheng Xu

Bin Ma

Horng Jyh Paul Wu

George M White

Rong Tong

Kai Xin

Gareth Loudon

Chitralekha Gupta

Disclosed is a system for assessing quality of a singing voice singing a song. The system comprises memory and at least one processor. The memory stores instructions that, when executed by the at least one processor, cause the at least one processor to receive a plurality of inputs comprising a first input and one or more further inputs, each input comprising a recording of a singing voice singing the song, to determine, for the first input, one or more relative measures of quality of the singing voice by comparing the first input to each further input; and to assess quality of the singing voice of the first input based on the one or more relative measures. Also disclosed is a method implemented on such a system.

System and method for assessing quality of a singing voice

A method for executing a fully mixed initiative dialogue (FMID) interaction between a human and a machine, a dialogue system for a FMID interaction between a human and a machine and a computer readable data storage medium having stored thereon computer code for instructing a computer processor to execute a method for executing a FMID interaction between a human and a machine are provided. The method includes retrieving a predefined grammar setting out parameters for the interaction; receiving a voice input; analyzing the grammar to dynamically derive one or more semantic combinations based on the parameters; obtaining semantic content by performing voice recognition on the voice input; and assigning the semantic content as fulfilling the one or more semantic combinations.

Dialogue system and a method for executing a fully mixed initiative dialogue (FMID) interaction between a human and a machine

A method for training a spoken language identification system to identify an unknown language as one of a plurality of known candidate languages includes the process of creating a sound inventory comprising a plurality of sound tokens, the collective plurality of sound tokens provided from a subset of the known candidate languages. The method further includes providing a plurality of training samples, each training sample composed within one of the known candidate languages. Further included is the process of generating one or more training vectors from each training database, wherein each training vector is defined as a function of said plurality of sound tokens provided from said subset of the known candidate languages. The method further includes associating each training vector with the candidate language of the corresponding training sample.

Spoken language identification system and methods for training and operating same

A method, an apparatus and a computer program product for adaptive, content-based watermark embedding of a digital audio signal ( ) are disclosed. Corresponding watermark extracting techniques are also disclosed. Watermark information ( ) is encrypted ( ) using an audio digest signal, i.e. a watermark key ( ). To optimally balance inaudibility and robustness when embedding and extracting watermarks ( ), the original audio signal ( ) is divided into fixed-length frames ( ) in the time domain. Echoes (S&prime;&lsqb;n&rsqb;, S&Prime;&lsqb;n&rsqb;) are embedded in the original audio signal ( ) to represent the watermark ( ). The watermark ( ) is generated by delaying and scaling the original audio signal ( ) and embedding it in the audio signal ( ). An embedding scheme ( ) is designed for each frame ( ) according to its properties in the frequency domain. Finally, a multiple-echo hopping module ( ) is used to embed and extract watermarks in the frame ( ) of the audio signal ( ). An audio watermarking system known as KentMark (Audio) is implemented.

Digital audio watermarking using content-adaptive, multiple echo hopping

A method of processing speech representative of ideograms for speech communication using an asynchronous communication channel ( ) is disclosed. The method includes the step of processing speech units of a speech and data indicative of the speech units. Each speech unit is representative of an ideogram or a plurality of semantically related ideograms ( ). The data indicative of the speech units is discretely communicable on the asynchronous communication channel ( ). By communicating the data indicative of the speech units, a substantially low data transmission rate and intelligible speech communication is achieved.

Low data transmission rate and intelligible speech communication

A method, an apparatus, a computer program product and a system for voice annotating and retrieving digital media content are disclosed. An annotation module ( ) post annotates digital media data ( ), including audio, image and/or video data, with speech. A word lattice ( ) can be created from speech annotation ( ) dependent upon acoustic and/or linguistic knowledge. An indexing module ( ) then indexes the speech-annotated data ( ). The word lattice ( ) is reverse indexed ( ), and content addressing ( ) is applied to produce the indexed data ( ). A speech query ( ) can be generated as input to a retrieval module ( ) for retrieving a segment of the indexed digital media data ( ). The speech query ( ) is converted into a word lattice ( ), and a shortlist ( ) is produced from it ( ) by confidence filtering ( ). The shortlist ( ) is input to a lattice search engine ( ) to search the indexed content ( ) to obtain the search result ( ).

Method and apparatus for voice annotation and retrieval of multimedia data

A system (,) for tokenization and named entity recognition of ideographic language is disclosed. In the system, a word lattice is generated for a string of ideographic characters using finite state grammars (,) and a system lexicon (,). Segmented text is generated by determining word boundaries in the string of ideographic characters using the word lattice dependent upon a contextual language model (,A) and one or more entity language models (,B). One or more named entities is recognized in the string of ideographic characters using the word lattice dependent upon the contextual language model (,A) and the one or more entity language models (,B). The contextual language model (,A) and the one or more entity language models (,B) are each class-based language models. The lexicon (,) includes single ideographic characters, words, and predetermined features of the characters and words.

Growing community of inventors

Singapore, Singapore

Haizhou Li