Nec Corporation

Seiya Shibata

Satoru Ishii

Hideo Hasegawa

Shintaro Nakano

Takashi Takenaka

Hiroaki Igarashi

Liang Feng

An object tracking device includes an object detection unit and a matching unit. The object detection unit includes a plurality of detectors having different processing speeds or accuracy. The object detection unit detects, for each of frames of image data input in time series, each object from the image data using one detector selected from the plurality of the detectors. The matching unit matches a current object detection result which the object detection unit generates from image data of a current frame with a previous object detection result which the object detection unit generates from image data of a previous frame before the current frame, and generates a tracking result of the object based on a matching result.

Object tracking device, object tracking method, and recording medium

An inferential device, including a quantization part that quantizes a result of a convolutional operation in a convolutional neural network using input data and weights; a convolutional operation part that performs a convolutional operation using the quantized operation result as input data; and an input data conversion part that converts the input data to a first layer to enable the convolutional operation part to process both the input data to the first layer and the input data that is quantized by the quantization part in a same way.

Inferential device, convolutional operation method, and program

An inference device comprises a weight storage part that stores weights, an input data storage part that stores input data, and a PE (Processing Element) that executes convolution computation in convolutional neural network using the weights and input data. The PE adds up weight elements to be multiplied with elements of the input data for each of variable values of the elements of the input data. The PE multiplies each of the variable values of the elements of the input data with each cumulative sum value of weights corresponding to the variable values of the elements of the input data. The PE adds up a plurality of multiplication results obtained by the multiplications.

Add-mulitply-add convolution computation for a convolutional neural network

A management apparatus () includes: a storage () that stores information indicating a correspondence between at least one virtual network function (VNF) operating on a server and a programmable logic circuit (FPGA) capable of operating at least part of a virtual network function; and a controller () that causes first and second servers to perform migration of a virtual network function operating on a programmable logic circuit of the first server to the second server, on the basis of the correspondence information.

Network system, management method and apparatus thereof, and server

An inference apparatus comprises a plurality of PEs (Processing Elements) and a control part. The control part operates a convolution operation in a convolutional neural network using each of a plurality of pieces of input data and a weight group including a plurality of weights corresponding to each of the plurality of pieces of input data by controlling the plurality of PEs. Further, each of the plurality of PEs executes a computation including multiplication of a single piece of the input data by a single weight and also executes multiplication included in the convolution operation using an element with a non-zero value included in each of the plurality of pieces of input data.

Inference apparatus, convolution operation execution method, and program

An information processing apparatus includes a sparse element detection part, a sparse location weight addition part, a multiplication part, a non-sparse data operation part, and an addition part. The sparse element detection part detects a predetermined sparse element from input data and outputs information about the sparse element. The sparse location weight addition part adds a first weight elements corresponding to the sparse element. The multiplication part multiplies an output of the sparse location weight addition part by the sparse element. The non-sparse data operation part performs an operation on non-sparse elements, each other than the sparse element in the input data. The addition part adds an output of the multiplication part and an output of the non-sparse data operation part.

Information processing apparatus, information processing method, and program

A network system includes multiple processing units (----) on each of which a desired virtual network function can be configured and a management apparatus that determines a communication path that connects the processing units so as to deploy a set of desired virtual network functions. At least one of the processing units includes a first communication interface that is connectable to any different processing unit and at least one second communication interface that is directly connectable to a predetermined different processing unit. The management apparatus determines the communication path for deploying the set of the desire virtual network functions, in accordance with respective connectable communication interfaces of the processing units.

Network system and management method and apparatus thereof

Provided is for reducing access latency. A high-level synthesis device includes feature quantity obtaining unit and implementation determination unit. Feature quantity obtaining unit obtains an access feature quantity including a feature quantity relating to communication between a plurality of modules by analyzing an access pattern in communication between the plurality of modules. Implementation determination unit determines an implementation method for communicating between the plurality of modules based on the obtained access feature quantity.

High-level synthesis device, high-level synthesis method, and program recording medium

An intra-prediction mode determination device is applied in a video coding device that recursively divides input video blocks into small blocks to perform coding by intra-prediction or inter-frame prediction, and includes a prediction mode selection unit and a number of bins adjustment unit that are provided in correspondence to only one of possible sizes of the small blocks. The prediction mode selection unit evaluates, with respect to each of the small blocks, the coding cost of a plurality of prediction mode candidates on the basis of a residual corresponding to the prediction mode candidates and the number of bins allocated to the prediction mode candidates, and selects an intra-prediction mode from the plurality of prediction mode candidates. The number of bins adjustment unit, when the prediction mode selection unit evaluates the coding cost, increases the number of bins corresponding to a specific prediction mode candidate.

Intra-prediction mode determination method, intra-prediction mode determination device, and storage medium for storing intra-prediction mode determination program

A processing control device includes: a process completion mapwhich is a map corresponding to processing units of respective sizes, and in which, when process corresponding thereto is completed, setting indicating completion of the process is performed; a dependence dissolution mapwhich is a map corresponding to the processing units of the respective sizes and in which information indicating that the dependence of each process disappears is set; a controllerwhich, when the process is completed, sets information indicating that the process is completed in an area corresponding to the process within a small size of map in the process completion map; and a reflecting circuitwhich reflects the information set in the process completion mapin the dependence dissolution map

Growing community of inventors

Tokyo, Japan

Seiya Shibata