Oki Electric Industry Co., Ltd.

Oki Data Corporation

Kurato Maeno

Kohei Yamamoto

Motoko Kagaya

Takashi Hashimoto

Michiyo Matsui

Akihiro Fujii

Daisuku Okuya

An information processing unit includes: an attention module having an attention layer and a computation section and that is for a neural network including plural levels of layer, the attention layer being configured to compute an output feature corresponding to an input feature from a predetermined layer and based on a parameter; the computation section that multiplies the input feature by the output feature, and outputs a computed result to a layer at a next level; the first learning unit connected to the neural network and that learns the parameter in a state in which learning has been suspended at least for the predetermined layer and the next level layer; and the channel selection section that selects as a redundant channel a channel satisfying a predetermined relationship between the output feature and a predetermined threshold value.

Neural network load reduction device, information processing unit, and neural network load reduction method and computer-readable storage medium

There is provided a neural network reduction device including: an input section configured to acquire a first neural network including a plurality of processing layers that includes at least one convolutional layer; a generation section configured to generate a third neural network in which at least one second neural network is connected as an additional layer to a stage subsequent to the convolutional layer included in the first neural network; a learning section configured to train the third neural network; a correction section configured to select, as a redundant channel, a channel whose feature value calculated on a basis of a weight parameter of the second neural network establishes a predetermined relationship with a predetermined threshold, delete the redundant channel from the convolutional layer, and correct a weight parameter of the convolutional layer after deletion of the redundant channel on a basis of at least a portion of the weight parameter of the second neural network; a relearning section configured to retrain the third neural network on a basis of correction of the weight parameter of the convolutional layer after the deletion of the redundant channel; and an output section configured to output the retrained third neural network.

Neural network reduction device, neural network reduction method, and storage medium

It is possible to improve estimation accuracy with regard to data in which significance is attached to a relative phase. Provided is an information processing device including an estimation unit configured to estimate a status by using a neural network. The neural network includes a first complex-valued neural network to which complex data is input, a phase difference computation layer from which phase difference for each element between a plurality of sets with regard to the complex data is output, and a second complex-valued neural network from which complex data is output on the basis of the phase difference.

Information processing device and information processing method

An estimation device includes a first converting unit configured to convert a beat signal to a one-dimensional first candidate signal on the basis of a two-dimensional distribution of the beat signal, a second converting unit configured to convert the beat signal to a one-dimensional second candidate signal on the basis of a two-dimensional position change of the beat signal, and a signal deciding unit configured to decide a one-dimensional signal on the basis of the first candidate signal and the second candidate signal.

Estimation device, vibration state estimation method, and recording medium

A tracking apparatus, including (1) a first target object tracking unit having a first processor, configured to correct a first information of at least one first candidate target object, detected by an image sensor, according to a first predetermined motion model, and to determine a distribution of the first information, (2) a second target object tracking unit including a second processor, configured to correct a second information of at least one second candidate target object, detected by an active type sensor, according to a second predetermined motion model, and to determine a distribution of the second information, (3) a matching unit configured to obtain a plurality of distribution parameters based on a correlation of the distributions of the first and second information, and (4) a false image determining unit configured to determine whether or not the second information corresponds to a false image based on the plurality of distribution parameters.

Tracking apparatus, tracking method, and computer-readable storage medium

Provided is a state recognizing device including: a first filter unit performing a filtering process on a time-series signal obtained by performing remote measuring on a space, in accordance with a first cutoff frequency to allow a signal component deriving from movement of a biological organism to pass through; a second filter unit performing a filtering process on the time-series signal in accordance with a second cutoff frequency to allow a signal component based on biological information to pass through; an amplitude-component extracting unit extracting a first feature quantity related to an amplitude component from a signal passed through the first and/or second filter units; a frequency-component extracting unit extracting a second feature quantity related to a frequency component from a signal passed through the first and/or second filter units; and a recognizing unit recognizing a state of the space by using the first and second feature quantities.

State recognizing device, state recognizing method, and recording medium

To improve the accuracy of the determination on whether or not a moving object is present in a predetermined region, a control device for use in a system configured to conduct presence-or-absence determination on whether or not a moving object is present on the basis of detection by a first sensor provided in a predetermined compartment region, the control device comprising: a mode control section configured to control switching of an operation mode for the detection by the first sensor or the presence-or-absence determination; and an acquisition section configured to acquire a determination result of whether or not a moving object is present in an adjacent region adjacent to the predetermined compartment region and not including the predetermined compartment region; wherein the mode control section controls the switching of the operation mode in response to the determination result acquired by the acquisition section, is provided.

Control device, and control system

The present invention provides an information printing method and a printing device, which assure a higher level of security when outputting printed matter. The present invention provides a printing device comprising a document storage means for storing electronic document data generated at a printing instruction issuing terminal operated by a printing instruction issuer and transmitted from the printing instruction issuing terminal, an ID setting means for setting a print retriever ID of a print retriever who is to execute processing for outputting the electronic document data stored at the document storage means onto paper, a watermark information superimposing means for superimposing the print retriever ID having been set at the ID setting means as watermark information onto the electronic document data and a printing means for outputting the electronic document data having the watermark information superimposed thereupon by the watermark information superimposing means onto paper.

Printing device and printing method generating watermark information

A microvibration feature quantity apparatus includes a beat signal receiver receiving a beat signal output from a Doppler sensor, which radiates a radiation wave at a predetermined frequency to a body to be observed and senses part of the radiation wave reflected by the body to output the beat signal at a frequency corresponding to a difference in frequency between the radiation wave and the reflected part of the wave, a frequency converter transforming the received beat signal into a frequency-domain signal, and a beat signal feature quantity calculator using the frequency-domain signal to derive a beat signal feature quantity exhibiting an abrupt change with respect to time of a vibration velocity of the microvibration, thereby accurately calculating a feature quantity of the microvibration raised in the body.

Apparatus and method for calculating microvibration feature quantity

In a watermarked document printing apparatus, secret information is divided, error correction encoding is performed to each divided pieces of information, and the secret information is converted into a pattern image and embedded in paper. In a watermarked document reading apparatus, after extracting signals from the pattern image on the paper, a reading error is detected by decoding the error correction code for each divided pieces of information, and the portion where the reading error exists is read again. The scanning process is performed again only to the portion where the decoding cannot be performed due to the temporal malfunction of the scanner or the like. Therefore, the pattern image can efficiently be obtained.

Watermarked document reading apparatus with improved error processing

In a copy management system, an output device generates a print image, and list information including identification information identifying a person authorized for printing. The list information is converted to a specific code, which is in turn embedded in a print image. The code-embedded print image is output by a printer on a printing medium. The code-embedded print image on the printing medium is received by a copying device of the system. The copying device also receives and accepts copier identification information identifying a copying person. The list information is acquired from the code-embedded print image. The identification information in the list information acquired is compared to the copier identification information accepted. Allowance/non-allowance of copying is determined based on whether or not the above information coincides with each other. The copying device outputs a printed copy subject to decision for allowance of copying.

System for managing flexible copying with information leakage prevented and/or detected

Printing device and printing method

The watermarked information embedding apparatus which inputs an image and embeds watermarked information in the input image, comprises: picture element determining means which determines whether it is a picture element constituting a background image for each of picture elements which constitute the input image; background picture element removing means which removes all of background picture elements determined as picture elements constituting the background image by the picture element determining means; and watermarked information embedding means which embeds the watermarked information in an image constituted by a picture element from which the background picture element constituting the input image is removed by the background picture element removing means.

Watermarked information embedding apparatus

The present invention is relate to watermark information embedding apparatus and method for the same, watermark information detecting apparatus and method for the same, and method of containing watermark document. A document image generator () generates a document image according to document data (). A watermark image generator () generates a watermark image. The watermark information () is denoted with dot pattern, and the dot pattern within the outline of the watermark information are of special value. A synthesizer () overlaps document image generated from the document image generator () and the watermark image generated from the watermark image generator () so as to generate a containing watermark document image.

Watermark information embedding device and method, watermark information detecting device and method, watermarked document

In a watermark processing system, a watermark detector detects first information represented by a watermark from an image, a watermark encoder decodes and error-corrects the first information and re-encodes it to second information, and a watermark eraser identifies the position of the tint-block pattern included in the watermark according to the second information and produces a watermark-erased image from which the watermark is erased.

Growing community of inventors

Saitama, Japan

Kurato Maeno