The invention relates to a support vector machine-based Vietnamese part-of-speech tagging method and belongs to the technical field of natural language processing. The method comprises the steps of firstly performing part-of-speech tagging modeling of Vietnamese words for a formed Vietnamese sentence-level part-of-speech tagging corpus to obtain a Vietnamese support vector machine part-of-speech tagging model; and randomly selecting a test corpus from a Vietnamese participle sentence-level corpus library, and performing part-of-speech tagging through the established Vietnamese support vector machine part-of-speech tagging model to obtain a parameter sequence result of part-of-speech tagging of the Vietnamese words. According to the method, the effective part-of-speech tagging of the Vietnamese words is realized and the powerful support is provided for work of lexical analysis, syntactic analysis, semantic analysis, information extraction, information retrieval, machine translation and the like; and the correct rate of part-of-speech tagging is high and a very good tagging effect is achieved.

Support vector machine-based Vietnamese part-of-speech tagging method

The objective is to limit recognition processing delay caused by the signal processing load. The components are a voice input means (10) that inputs a voice signal, an output judgment means (20) that determines if the voice signal input by the voice input means (10) is a voice signal in a sound segment, which is a voice segment in which a speaker is considered to be speaking, and that outputs said voice signal for processing only when it is determined to be a voice signal in a sound segment, a signal processing means (30) that performs signal processing for the voice signal output as a voice signal for processing by the output judgment means (20), and a speech recognition processing means (40) that performs speech recognition processing for the voice signal processed by the signal processing means (30).

Speech recognition system, speech recognition method, and storage medium where speech recognition program is stored

Provided in the embodiments of the present invention are a signal processing method and apparatus, comprising: collecting and detecting voice signals by means of a microphone; by means of a target network model, identifying the probability of the detected voice signals belonging to each type of signal; and classifying the detected voice signals into signal types that have the greatest identified probability. By means of the embodiment of the present invention, voice diseases may be examined by means of a non-contact manner.

Signal processing method and apparatus

A purpose is to suppress recognition process delay generated due to load in signal processing. Included is a speech input means 10 that inputs a speech signal, an output evaluation means 20 that evaluates whether or not the speech signal input by the speech input means 10 is the speech signal in a sound section, which is a speech section assuming that a speaker is speaking, and outputs the speech signal as a speech signal to be processed only when evaluated as the speech signal in the sound section, a signal processing means 30 that performs signal processing to the speech signal, which is output by the output evaluation means 20 as the speech signal to be processed, and a speech recognition processing means 40 that performs a speech recognition process to the speech signal which is signal-processed by the signal processing means 30.

Speech recognition system to evaluate speech signals, method thereof, and storage medium storing the program for speech recognition to evaluate speech signals

The embodiment of the invention provides a signal processing method and device. The signal processing method comprises the steps that acceleration signals of the skin surface of the sound production part of a person are acquired; tone signals are extracted from the acceleration signals; tone characteristics are extracted from the tone signals; the tone characteristics are classified. By adopting the signal processing method and device, the tone classification accuracy can be improved.

Signal processing method and device

PROBLEM TO BE SOLVED: To provide a speech recognizer that has less throughput and high recognition performance especially for speech recognition of a tone language.SOLUTION: The speech recognizer: extracts a fundamental frequency from input speech, and acoustically analyses the input speech; selects one of plural speech recognition results obtained by speech recognition, and outputs a label string indicating the selected speech recognition result; selects at least one label in the output label string on the basis of a pattern list held in advance; and extracts tone information indicating tone of the selected label on the basis of the fundamental frequency extracted from the input speech, and corrects the selected label on the basis of the extracted tone information and content of the pattern list.

Speech recognizer and speech recognition method

PROBLEM TO BE SOLVED: To generate no deterioration of tone quality even in voice encoding encoding exceeding a frame cycle by eliminating a part of a digital voice data collected at the next frame timing and successively collecting a digital voice data when an encoding processing time exceeds the frame. SOLUTION: In the case that frame-data-collection completion information is inputted to a frame synchronization detection part 108 before encoding completion information, an elimination instruction is outputted to a sample data elimination part 104. The sample elimination part 104 reads the digital voice data from a frame data buffer 102 and a pitch cycle of the immediately preceding frame from a pitch extract part 103 respectively, and deletes the digital voice data for a pitch at that pitch cycle. Further, the number of the deleted samples is outputted to a frame sample count part 110 as an eliminated count number. The eliminated frame voice data are inputted to a frame sample data connection part 111, and the sample data connected with the eliminated part inputted to the frame buffer 102.

Shosakai Makoto

Papers

Speech recognizer and speech recognition method

Voice encoding method