Sony Patent | Information processing device, information processing method, and computer program

Patent: Information processing device, information processing method, and computer program

Publication Number: 10175765

Publication Date: 2019-01-08

Applicants: Sony

Abstract

An information processing device that detects the position of a subject shot by a camera in a three-dimensional space is provided. The information processing device includes a depth image acquiring section that acquires a depth image in which the distance of a subject in the depth direction from the camera is represented as a pixel value on an image plane, a spatial structure identifying section that identifies the position of a plane forming a predetermined angle with a gravity vector among subject surfaces by detecting a coordinate point collection representing the plane on the basis of the distribution of coordinate points obtained by inverse projection of pixels of the depth image to the three-dimensional space based on the direction of the gravity vector, and a target object extracting section that generates and outputs position information of a predetermined target object on the basis of the position of the plane.

Background

The present disclosure relates to an information processing device that executes processing on the basis of a shot image, an information processing method carried out by the information processing device, and computer program.

In recent years, it has been becoming common to equip a personal computer, a game machine, etc. with a camera and shoot a figure of a user to use the shot image in various forms. For example, systems to transmit an image of a user to the other side as it is via a network, such as a video phone and a video chat, and systems to recognize the motion of a user by image analysis and use the recognized motion as input information for a game or information processing have been put into practical use (refer to e.g. WO 2007/050885 A2 bulletin). Moreover, in recent years, it has been becoming possible to realize games and image expression with a more realistic feeling by accurately detecting the motion of a target object in a three-dimensional space including the depth direction.

Summary

Various problems exist in shooting a space in which a wide variety of objects exists to determine only a desired target object from the shot image and track the motion of the desired target object. For example, in a general position detection technique using a stereo image method or infrared reflection, images in a shot image are captured as collections of discrete points and the positions of them in a three-dimensional space are acquired. Therefore, in some cases, it is difficult to accurately acquire geometric, structural information such as information on which direction and what positional relationship an objects exist in. This often results in failure in acquisition of sufficient accuracy to execute information processing according to the motion of a person in contact with a floor or a wall or implement augmented reality (AR), in which image processing is so carried out that a three-dimensional object rendered by computer graphics appears to exist on a floor or a table as an actually-shot image.

There is a need for the present disclosure to provide a technique by which position detection of a target object in a three-dimensional space can be carried out with high accuracy.

According to an embodiment of the present disclosure, there is provided an information processing device that detects a position of a subject shot by a camera in a three-dimensional space. This information processing device includes a depth image acquiring section configured to acquire a depth image in which distance of a subject in depth direction from the camera is represented as a pixel value on an image plane, a spatial structure identifying section configured to identify a position of a plane forming a predetermined angle with a gravity vector among subject surfaces by detecting a coordinate point collection representing the plane on a basis of distribution of coordinate points obtained by inverse projection of pixels of the depth image to the three-dimensional space based on direction of the gravity vector, and a target object extracting section configured to generate and output position information of a predetermined target object on a basis of the position of the plane.

According to another embodiment of the present disclosure, there is provided an information processing method in which an information processing device detects a position of a subject shot by a camera in a three-dimensional space. This information processing method includes acquiring a depth image in which distance of a subject in depth direction from the camera is represented as a pixel value on an image plane and storing the depth image in a memory, identifying a position of a plane forming a predetermined angle with a gravity vector among subject surfaces by detecting a coordinate point collection representing the plane on a basis of distribution of coordinate points obtained by inverse projection of pixels of the depth image read from the memory to the three-dimensional space based on direction of the gravity vector, and generating and outputting position information of a predetermined target object on a basis of the position of the plane.

According to further another embodiment of the present disclosure, there is provided a computer program that causes a computer to implement detecting a position of a subject shot by a camera in a three-dimensional space. The computer program causing the computer to implement a configuration includes: acquiring a depth image in which distance of a subject in depth direction from the camera is represented as a pixel value on an image plane; identifying a position of a plane forming a predetermined angle with a gravity vector among subject surfaces by detecting a coordinate point collection representing the plane on a basis of distribution of coordinate points obtained by inverse projection of pixels of the depth image to the three-dimensional space based on direction of the gravity vector; and generating and outputting position information of a predetermined target object on a basis of the position of the plane.

What are obtained by translating arbitrary combinations of the above constituent elements and expressions of the present disclosure among method, device, system, computer program, recording medium in which a computer program is recorded, and so forth are also effective as embodiments of the present disclosure.

According to the embodiments of the present disclosure, information processing based on the position and motion of a target object in a shot image can be implemented with high accuracy.

发表评论

电子邮件地址不会被公开。 必填项已用*标注