Kabushiki Kaisha Toshiba

Toshiba Medical Systems Corporation

Takeshi Nagai

Takeshi Chujoh

Yoshihiro Kikuchi

Shinichiro Koto

Wataru Asano

Tomoo Yamakage

Akiyuki Tanizawa

Tomoya Kodama

Nakaba Kogure

Noboru Yamaguchi

Shuhei Nitta

Yukinobu Sakata

Koichi Masukura

Atsushi Yaguchi

Naoto Date

Taku Kato

Toshimitsu Kaneko

Hidenori Takeshima

Atsushi Ishihara

Takeshi Mita

Hiroaki Nakai

Keiri Nakanishi

Ryusuke Hirai

Takayuki Itoh

Tadaaki Masuda

Shunichi Gondo

Yasuki Tanabe

Kazuki Inoue

Takuya Matsuo

A learning device includes a structure search unit that searches for a first learned model structure obtained by selecting search space information in accordance with a target constraint condition of target hardware for each of a plurality of convolution processing blocks included in a base model structure in a neural network model; a parameter search unit that searches for a learning parameter of the neural network model in accordance with the target constraint condition; and a pruning unit that deletes a unit of at least one of the plurality of convolution processing blocks in the first learned model structure based on the target constraint condition and generates a second learned model structure.

Learning device, learning system, and learning method

A learning device according to an embodiment includes one or more hardware processors configured to function as a generation unit, an inference unit, and a training unit. The generation unit generates input data with which an error between a value output from each of one or more target nodes and a preset aimed value is equal to or less than a preset value, the target nodes being in a target layer of a plurality of layers included in a first neural network. The inference unit causes the input data to propagate in a forward direction of the first neural network to generate output data. The training unit trains a second neural network differing from the first neural network by using training data including a set of the input data and the output data.

Learning device, learning method, and computer program product

According to an embodiment, an information processing apparatus includes a computing unit and a compressing unit. The computing unit is configured to execute computation of an input layer, a hidden layer, and an output layer of a neural network. The compressing unit is configured to irreversibly compress output data of at least a part of the input layer, the hidden layer, and the output layer and output the compressed data.

Irreversible compression of neural network output

According to an embodiment, a video compression apparatus includes a first compressor, a second compressor, a partitioner and a communicator. The first compressor compresses a first video to generate a first bitstream. The second compressor sets regions in a second video and compresses the regions so as to enable each region to be independently decoded, to generate a second bitstream. The partitioner partitions the second bitstream according to the set regions to obtain a partitioned second bitstream. The communicator receives region information indicating a specific region that corresponds to one or more regions and selects and transmits a bitstream corresponding to the specific region from the partitioned second bitstream.

Video compression apparatus and video playback apparatus

A data compressing device according to an embodiment includes a data cutting unit configured to divide continuously inputted data into W-bit data blocks and to output the data blocks in segments such that each of the segments is composed of N data blocks, and a compression-method determining unit configured to select, as a compression portion for each of the segments, a run length system, a flag system, or no compression, according to a ratio of data blocks of specific data in any of the segments. The data compressing device further includes an RL compression unit configured to execute, on any of the segments, a run length system of storing a consecutive amount of the specific data into compressed data, and a flag compression unit configured to execute, on any of the segments, a flag system of storing positional information of the specific data into compressed data.

Data compressing device, data decompressing device, and data compressing/decompressing apparatus

A medical image processing apparatus according to an embodiment includes compressing circuitry. The compressing circuitry is configured to compress, for each of a first data group and a second data group both of which being included in data pertaining to a medical image, each of the first data group and the second data group separately, starting from data corresponding to a boundary between the first data group and the second data group and shifting sequentially to a direction away from the boundary.

Medical image processing apparatus and X-ray CT apparatus

In an image coding apparatus, a transformation unit acquires a transformation coefficient by executing frequency transformation to an image. An extraction unit extracts a plurality of bits each from the transformation coefficient. A coding unit executes variable-length coding to the plurality of bits including a most significant bit, and executes fixed-length coding to the plurality of bits including a least significant bit. In an image decoding apparatus, an extraction unit extracts a plurality of bits from a code string. A decoding unit executes variable-length decoding to the plurality of bits including a most significant bit, and executes fixed-length decoding to the plurality of bits including a least significant bit. A connection unit acquires a transformation coefficient by connecting the plurality of bits each decoded. An inverse transformation unit generates the image by executing frequency-inverse transformation to the transformation coefficient.

Apparatus for coding or decoding an image by executing variable-length coding/decoding and fixed-length coding/decoding

A video decoding apparatus includes a decoder to decode video encoded data using a predictive picture signal for a video signal, a compression device compressing the decoded picture signal, a first memory storing the decoded picture signal, a second memory storing the compressed picture signal, a decompression device decompressing the compressed picture signal read from the second memory, a selector selecting one of a decoded picture signal read from the first memory and a compressed/decompressed picture signal from the compression device as a reference picture signal according to at least one of a coding type of the video encoded data in picture unit and a prediction mode in block unit, and a motion compensator performing motion compensation on the reference picture signal to generate a predictive picture signal.

Video decoding method and apparatus

A fade detection unit performs fade detection, and in a case where a judgment of fade-in has been made, an encoding control unit notifies a motion detection unit to perform encoding in a reverse order to a display order. The fade detection unit performs the fade detection, and in the case where the judgment of the fade-in or fade-out has been made, the encoding control unit notifies the motion detection unit to perform encoding while importance is attached to a prediction from an image having a larger information amount in a B picture.

Dynamic image encoding device and method

A motion picture encoding method includes searching for a first motion vector between an object block of an input motion picture and a reference image for a frame-based prediction mode, searching for a second motion vector between the object block and the reference image for a field-based prediction mode, generating a first and second predictive residuals in accordance with the first and second motion vector, extracting a first and second feature quantities indicating a first and second similarity between lines of the first and second predictive residual, selecting the frame-based prediction mode in a case where the first similarity is larger than the second similarity based on the first and the second feature quantity, selecting the field-based prediction mode in a case where the second similarity is larger than the first similarity, and inter-frame-coding the input motion picture in accordance with the selected prediction mode.

Motion picture encoding method and device

When a search is made for a motion vector of a moving image, a motion vector search desirable against a hard movement is devised to be made with a small processing amount, and a motion vector detection device includes a wide range motion vector search part, a narrow range motion vector search part, a motion vector prediction part, and a motion vector re-search part. A rough search is made at equal intervals on an inputted image while a corresponding position in the reference image is made the center; a fine search is made in the vicinity of a predicted value of the motion vector only at one stage in accordance with a search format of a gradient method; and from a result of both, motion vector detection is performed as a motion vector search start point by the gradient method.

Motion vector detection method and device of the same

A digital watermark embedding apparatus includes a detector detecting a signal characteristic of at least one of an input image signal and the output image signal, and a controller controlling an embedding intensity of a digital watermark signal with respect to the input image signal in accordance with the signal characteristic.

Digital watermark embedding apparatus and digital watermark detecting apparatus

An image size reduction unit reduces the image size of an input image signal, thereby generating a size-reduced image signal. A specific frequency component extraction unit extracts a specific frequency component signal from the size-reduced image signal. A phase controller controls the phase of the specific frequency component signal. A correlator computes a cross-correlation value between the phase-controlled specific frequency component signal and size-reduced image signal. A watermark information estimation unit estimates, from the cross-correlation value, watermark information embedded in the input image signal.

Digital watermark detection method and apparatus

A specific frequency component extraction unit extracts a specific frequency component signal from an input image signal. A phase controller controls the phase of the specific frequency component signal. A correlator computes the cross-correlation value of the phase-controlled specific frequency component signal and the input image signal. A watermark information estimation unit filters the cross-correlation value to detect watermark information embedded in the input image signal.

Growing community of inventors

Yokohama, Japan

Wataru Asano