Information processing device, control method, and program

Patent metadata

Field	Value
Publication number	US-12039451-B2
Application number	US-201817059678-A
Country	US
Kind code	B2
Filing date	Jun 1, 2018
Priority date	Jun 1, 2018
Publication date	Jul 16, 2024
Grant date	Jul 16, 2024

Abstract

An information processing apparatus ( 2000 ) generates likelihood data for each of a plurality of partial regions ( 12 ) in image data ( 10 ). The likelihood data are data being associated with a position and a size on the image data ( 10 ) and indicating a likelihood that a target object exists in an image region at the position with the size. The information processing apparatus ( 2000 ) computes a distribution (probability hypothesis density: PHD) of an existence likelihood of a target object with respect to a position and a size by computing the total sum of likelihood data each piece of which is generated for each partial region ( 12 ). The information processing apparatus ( 2000 ) extracts, from the PHD, partial distributions each of which relates to one target object. For each extracted partial distribution, the information processing apparatus ( 2000 ) outputs a position and a size of a target object represented by the partial distribution, based on a statistic of the partial distribution.

First claim

The invention claimed is:

1. An information processing apparatus comprising:
at least one memory configured to store a computer program; and
at least one processor configured to execute the computer program to perform:
acquiring image data and generating likelihood data representing a likelihood of existence of a target object with respect to a position and a size for each of a plurality of partial regions included in the image data;
computing a distribution of the likelihood of existence of the target object with respect to the position and the size by computing a total sum of the likelihood data each piece of which is generated for the each partial region and extracting, from the computed distribution, one or more partial distributions each of which relates to the one target object; and
outputting, for the each extracted partial distribution, the position and the size of the target object relating to the partial distribution, based on a statistic of the partial distribution,
wherein with respect to the position and the size of the partial region, the likelihood data represent a distribution conforming to a predetermined model and indicating a likelihood that an object exists in the partial region with the position and the size,
the computer program includes a neural network outputting the likelihood data for each of a plurality of partial regions included in image data and,
the generating the likelihood data includes, by inputting the acquired image data to the neural network, generating the likelihood data for each of the plurality of partial regions included in the image data,
for the each partial region, the neural network outputs the likelihood that the target object exists in the partial region and a parameter value of the predetermined model.

2. The information processing apparatus according to claim 1 , wherein the extracting the one or more partial distributions includes;
computing a number of one or more objects included in the image data, based on an integral value of a distribution represented by a total sum of the likelihood data, and
extracting as many as the number of the partial distributions from the distribution represented by the total sum of the likelihood data.

3. The information processing apparatus according to claim 1 , wherein the extracting the one or more partial distributions includes extracting the partial distributions an integral value of each of which is 1 from a distribution represented by a total sum of the likelihood data.

4. The information processing apparatus according to claim 1 , wherein the generating likelihood data includes generating the likelihood data for each of mutually of mutually different types of the target objects,
the computing the distribution of the likelihood includes, for each of mutually different types of the target objects, computing the distribution of the likelihood of existence of the target object,
the one or more partial distributions is extracted from the distribution, and
the outputting the position and the size of the target object includes outputting the position and the size of the target object relating to the each partial distribution along with a type of the target object relating to the partial distribution.

5. A control method executed by a computer, the control method comprising:
acquiring image data and generating likelihood data representing a likelihood of existence of a target object with respect to a position and a size for each of a plurality of partial regions included in the image data;
computing a distribution of the likelihood of existence of the target object with respect to the position and the size by computing a total sum of likelihood data each piece of which is generated for the each partial region and extracting, from the computed distribution, one or more partial distributions each of which relates to the one target object; and
outputting, for the each extracted partial distribution, the position and the size of the target object relating to the partial distribution, based on a statistic of the partial distribution,
wherein with respect to the position and the size of the partial region, the likelihood data represent a distribution conforming to a predetermined model and indicating a likelihood that an object exists in the partial region with the position and the size,
a neural network is stored in the computer, outputting the likelihood data for each of a plurality of partial regions included in image data and,
the generating the likelihood data includes, by inputting the acquired image data to the neural network, generating the likelihood data for each of the plurality of partial regions included in the image data,
for the each partial region, the neural network outputs the likelihood that the target object exists in the partial region and a parameter value of the predetermined model.

6. The control method according to claim 5 , wherein the extracting the one or more partial distributions includes;
computing a number of one or more objects included in the image data, based on an integral value of a distribution represented by a total sum of the likelihood data, and
extracting as many as the number of the partial distributions from the distribution represented by the total sum of the likelihood data.

7. The control method according to claim 5 , wherein the extracting the one or more partial distributions includes extracting the partial distributions an integral value of each of which is 1 from a distribution represented by a total sum of the likelihood data.

8. The control method according to claim 5 , wherein the generating likelihood data includes generating the likelihood data for each of mutually different types of the target objects,
the computing the distribution of the likelihood includes, for each of mutually different types of the target objects, computing the distribution of the likelihood of existence of the target object,
the one or more partial distributions is extracted, and
the outputting the position and the size of the target object includes outputting the position and the size of the target object relating to the each partial distribution along with a type of the target object relating to the partial distribution.

9. A non transitory storage medium storing a program causing a computer to execute:
acquiring image data and generating likelihood data representing a likelihood of existence of a target object with respect to a position and a size for each of a plurality of partial regions included in the image data;
computing a distribution of the likelihood of existence of the target object with respect to the position and the size by computing a total sum of likelihood data each piece of which is generated for the each partial region and extracting, from the computed distribution, one or more partial distributions each of which relates to the one target object; and
outputting, for the each extracted partial distribution, the position and the size of the target object relating to the partial distribution, based on a statistic of the partial distribution,
wherein with respect to the position and the size of the partial region, the likelihood data represent a distribution conforming to a predetermined model and indicating a likelihood that an object exists in the partial region with the position and the size,
a neural network is included in the program, outputting the likelihood data for each of a plurality of partial regions included in image data and,
the generating the likelihood data includes, by inputting the acquired image data to the neural network, generating the likelihood data for each of the plurality of partial regions included in the image data,
for the each partial region, the neur

Assignees

• Nec Corp

Inventors

• Miyano Hiroyoshi
• Suzuki Tetsuaki

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06V40/10 Primary

  Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06F18/2415

  based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate · CPC title