20230095977. INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND PROGRAM simplified abstract (Sony Interactive Entertainment Inc.)

From WikiPatents
Jump to navigation Jump to search

INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND PROGRAM

Organization Name

Sony Interactive Entertainment Inc.

Inventor(s)

Yoshikazu Onuki of Tokyo (JP)

INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND PROGRAM - A simplified explanation of the abstract

This abstract first appeared for US patent application 20230095977 titled 'INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND PROGRAM

Simplified Explanation

The patent application describes an information processing apparatus that uses generative adversarial networks (GANs) to generate image data in different domains. Specifically, it focuses on generating image data of a first domain from image data of a second domain and vice versa.

  • The apparatus includes a generator building section that constructs a generator capable of generating image data of the first domain from image data of the second domain.
  • The generator also has the ability to generate image data of the second domain from the image data of the first domain.
  • The first domain consists of image data with multiple channels, including two time series images. Each time series image contains two objects that tend to move differently.
  • The second domain is defined by image data with multiple channels, including the same two time series images but excluding the second object.

Potential applications of this technology:

  • Image translation: The apparatus can be used to convert image data from one domain to another, allowing for the translation of images with different objects and movements.
  • Data augmentation: By generating image data in different domains, the apparatus can be used to expand training datasets for machine learning models, improving their performance.
  • Video editing: The technology can assist in modifying and enhancing videos by generating new image data in different domains, providing creative possibilities for video editing.

Problems solved by this technology:

  • Limited training data: The apparatus addresses the issue of limited training data by generating additional image data in different domains, allowing for more diverse and comprehensive training.
  • Object-specific image generation: The technology enables the generation of image data with specific objects and their distinct movements, which can be useful in various applications such as computer vision and robotics.

Benefits of this technology:

  • Improved image translation: The use of GANs and the generator building section allows for more accurate and realistic image translation between different domains.
  • Enhanced data diversity: By generating image data in different domains, the technology increases the diversity and richness of available training data, leading to better machine learning models.
  • Flexibility in object manipulation: The apparatus provides the ability to manipulate and control specific objects and their movements in image data, offering new possibilities for various applications.


Original Abstract Submitted

there is provided an information processing apparatus including a generator building section configured to build a generator for generating image data of a first domain from image data of a second domain and for generating the image data of the second domain from the image data of the first domain by use of cycle gan (generative adversarial networks). the first domain is defined by image data of a plurality of channels including at least two time series images each including a first object and a second object, each of the objects having a tendency to move differently. the second domain is defined by the image data of a plurality of the channels including at least the two time series images each including the first object and excluding the second object.