Nec Corporation

Ken Hanazawa

Hiroshi Hashimoto

Kan Arai

Makiko Akiguchi

Azusa Furukawa

Kei Shibuya

Masanori Tsujikawa

Akihiko Sugiyama

Takayuki Arakawa

Koji Okabe

Daisuke Tanaka

Ryosuke Isotani

Seiya Osada

Hitoshi Yamamoto

Kiyoshi Yamabana

Kiyokazu Miki

Shuji Komeiji

Fumihiro Adachi

In a familiarity degree estimation apparatus, a first video process unit calculates a time while a hand being viewed, which is a time when a customer is viewing the hand. A second video process unit calculates a time while an item being held, which is a time when the customer is holding an item. A familiarity degree estimation unit calculates a time while the item being viewed based on the time while a hand being viewed and the time while the item being held, and estimates that the longer the time while the item being viewed, the lower the degree of familiarity of the customer with respect to the item.

Familiarity degree estimation apparatus, familiarity degree estimation method, and recording medium

A customer analysis apparatus () includes an image acquisition unit () and an image processing unit (). The image acquisition unit () acquires a first image generated by a first image capture unit () and a second image generated by a second image capture unit (). The image processing unit () processes the first image and the second image. More specifically, the image processing unit () generates product information indicating a product () taken out from a product shelf (), by processing the first image and the second image. Further, the image processing unit () generates customer attribute information of a customer taking out the product (), by processing the first image. Then, the image processing unit () causes a result storage unit () to store the product information and the customer attribute information in association with each other.

Customer analysis apparatus, customer analysis method, and non-transitory storage medium

The present invention provides a processing apparatus () including: a classification execution unit () that executes class classification, based on an inference device; a computation unit () that computes a precision of a class determined by the class classification; a processing unit () that outputs a class determined by the class classification and accepts a correct/incorrect input from an operator when the precision is less than a reference value; and a registration unit () that registers a product determined by the class classification as a purchase target when the precision is equal to or more than the reference value, and registers a product determined by a correct/incorrect input from the operator as a purchase target when the precision is less than the reference value.

Receive operator input correcting product information when precision of identification of product is less than reference value

The present invention provides a processing apparatus () including: a registration unit () that recognizes, based on image processing, product discrimination information of a product picked up from a product placing space, and registers the recognized product discrimination information in association with customer discrimination information; and a settlement unit () that executes, when it is detected, based on image processing, that a customer having picked up the product performs a predetermined action, a settlement based on a registration content resulting from the registration unit ().

Processing apparatus, processing method, and non-transitory storage medium

An apparatus of this invention is directed to a speech input apparatus for obtaining a piece of high-quality speech with included noise reduced. The speech input apparatus includes a sound insulating block attached to a ceiling member in a vehicle, a first microphone that is attached to a side surface of the sound insulating block or the ceiling member in the vicinity of the sound insulating block on a side close to a passenger of the vehicle, inputs a sound mixture including the voice of the passenger and noise in the vehicle, and outputs a first signal to a noise suppressor configured to output an enhanced speech signal, and a second microphone that is attached to a side surface of the sound insulating block or the ceiling member in the vicinity of the sound insulating block on a side far from the passenger, inputs the noise in the vehicle, and outputs a second signal to the noise suppressor.

Speech input apparatus, speech processing method, speech processing program, ceiling member, and vehicle

A system of this invention is directed to a speech processing system that efficiently performs noise suppression processing for a plurality of noise sources spreading in a lateral direction with respect to a speaker of interest. The speech processing system includes a microphone array including a plurality of microphones, each of which inputs a sound mixture including speech of a speaker of interest and noise from a noise source region including a plurality of noise sources placed in a lateral direction with respect to the speaker of interest, and outputs a mixture signal including a speech signal and a noise signal, the plurality of microphones being arranged such that a difference between respective distances from the plurality of microphones to the speaker of interest becomes different from a difference between respective distances from the plurality of microphones to the noise source region, and a noise suppressor that suppresses the noise based on the mixture signals output from the plurality of microphones.

Speech processing system, speech processing method, speech processing program, vehicle including speech processing system on board, and microphone placing method

To input the voice of the passenger of a vehicle and output a piece of high-quality enhanced speech independently of the direction of a piece of speech or noise, a speech processing apparatus includes a first microphone that is provided on one of a ceiling member in a vehicle and an accessory thereof, inputs a sound mixture including a voice of a passenger of the vehicle and noise in the vehicle, and outputs a first signal, a second microphone that is provided on one of the ceiling member in the vehicle and the accessory thereof at a position farther than the first microphone when viewed from the passenger of the vehicle, inputs the noise in the vehicle while insulating the voice of the passenger of the vehicle using one of the ceiling member of the vehicle and the accessory thereof, and outputs a second signal, and a noise suppressor that outputs an enhanced speech signal based on the first signal and the second signal.

Speech processing apparatus, speech processing method, speech processing program, method of attaching speech processing apparatus, ceiling member, and vehicle

A speech recognition unit () includes a phrase determination unit () which determines a phrase boundary based on the comparison between the hypothetical word group generated by speech recognition and set words representing phrase boundaries. In this speech processing device, the speech recognition unit () outputs recognition results for each phrase based on a phrase boundary determined by the phrase determination unit ().

Speech processing device, method, and storage medium

Provided are a noise reduction system that highly precisely estimates noise contained in an input signal and highly precisely reduces the noise contained in the input signal using the estimated noise, a speech detection system, a speech recognition system, a noise reduction method, and a noise reduction program. The noise reduction system includes: a first noise estimating unit () that estimates a stationary noise component contained in a first input signal; a first noise reduction unit () that reduces the stationary noise component from the first input signal; a second noise estimating unit () that re-estimates a stationary noise component contained in the first input signal; a third noise estimating unit () that estimates a second non-stationary noise component including a sum of a stationary noise component and a non-stationary noise component contained in the first input signal; an estimated noise combining unit () that estimates a stationary noise component and a second non-stationary noise component contained in the first input signal; and a second noise reduction unit () that reduces the stationary noise component and the second non-stationary noise component contained in the first input signal.

Noise reduction system, speech detection system, speech recognition system, noise reduction method, and noise reduction program

A plurality of pruning measures (PM) are calculated from a feature amount (CV) of test data (TD) which is input, a plurality of isopycnic surfaces (EC) are plotted and set on a threshold space (SS), a threshold curved surface (SC) in which a decrease in at least one of a plurality of pruning measures (PM) causes an increase in at least one thereof is generated using a portion of one isopycnic surface (EC) as a part, a hypothesis curved surface (HC) of subject data (CD) is generated on the threshold space (SS) to set a position intersecting the threshold curved surface (SC) to a pruning threshold (PS), and a plurality of hypotheses of the subject data (CD) are pruned. Thereby, there is provided a data processing device of which at least one of the recognition speed and the recognition accuracy is higher than in the related art.

Data processing device, information storage medium storing computer program therefor and data processing method

This devicestores a first content-specific language model representing a probability that a specific word appears in a word sequence representing a first content, and a second content-specific language model representing a probability that the specific word appears in a word sequence representing a second content. Based on a first probability parameter representing a probability that a content represented by a target word sequence included in a speech recognition hypothesis generated by a speech recognition process of recognizing a word sequence corresponding to a speech, a second probability parameter representing a probability that the content represented by the target word sequence is a second content, the first content-specific language model and the second content-specific language model, the device creates a language model representing a probability that the specific word appears in a word sequence corresponding to a part corresponding to the target word sequence of the speech.

Language model creation device

A speech recognition apparatus that performs frame synchronous beam search by using a language model score look-ahead value prevents the pruning of a correct answer hypothesis while suppressing an increase in the number of hypotheses. A language model score look-ahead value imparting deviceis provided with a word dictionarythat defines a phoneme string of a word, a language modelthat imparts a score of appearance easiness of a word, and a smoothing language model score look-ahead value calculation means. The smoothing language model score look-ahead value calculation meansobtains a language model score look-ahead value at each phoneme in the word from the phoneme string of the word defined by the word dictionaryand the language model score defined by the language modelso that the language model score look-ahead values are prevented from concentrating on the beginning of the word.

Language model score look-ahead value imparting device, language model score look-ahead value imparting method, and program storage medium

The present invention enables the recognition process at high speed even when a lot of garbage is included in the grammar. The first voice recognition processing unit generates a recognition hypothesis graph which indicates a structure of hypothesis that is derived according to a first grammar together with a score associated with respective connections of a recognition unit by executing a voice recognition process based on the first grammar to a voice feature amount of input voice, and the second voice recognition processing unit outputs the recognition result from a total score of a hypothesis which is derived according to a second grammar after executing a voice recognition process according to the second grammar that is specified to accept a section other than keywords in input voice as the garbage section to a voice feature amount of input voice, and the second voice recognition processing unit acquires the structure and the score of the garbage section from the recognition hypothesis graph.

Growing community of inventors

Tokyo, Japan

Ken Hanazawa