18102265. METHOD AND APPARATUS WITH IMAGE RECONSTRUCTION simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)

From WikiPatents
Jump to navigation Jump to search

METHOD AND APPARATUS WITH IMAGE RECONSTRUCTION

Organization Name

SAMSUNG ELECTRONICS CO., LTD.

Inventor(s)

INWOO Ha of Suwon-si (KR)

METHOD AND APPARATUS WITH IMAGE RECONSTRUCTION - A simplified explanation of the abstract

This abstract first appeared for US patent application 18102265 titled 'METHOD AND APPARATUS WITH IMAGE RECONSTRUCTION

Simplified Explanation

An image reconstruction method and apparatus are disclosed in this patent application. The method involves several steps to reconstruct an image using neural network models and image warping techniques. Here is a simplified explanation of the abstract:

  • The method starts by warping a previous reconstruction result using change-data that represents the difference between rendered images.
  • A previous filter kernel is determined by executing a first neural network model with a previous rendered image and the image warping result.
  • The previous filter kernel is then warped using the change-data to estimate a current filter kernel.
  • Finally, a second neural network model is executed with a current rendered image, the current filter kernel, and the image warping result to determine the current reconstruction result.

Potential applications of this technology:

  • Image reconstruction in computer graphics and animation: This method can be used to improve the quality of reconstructed images in computer graphics and animation applications.
  • Image super-resolution: By reconstructing images using neural network models and image warping, this technology can enhance the resolution of low-resolution images.

Problems solved by this technology:

  • Image artifacts: The use of image warping and neural network models helps to reduce artifacts and improve the quality of reconstructed images.
  • Inaccurate reconstruction: By incorporating change-data and previous filter kernels, this method improves the accuracy of image reconstruction compared to traditional methods.

Benefits of this technology:

  • Improved image quality: The combination of image warping and neural network models results in higher-quality reconstructed images.
  • Enhanced efficiency: By utilizing previous reconstruction results and filter kernels, this method reduces the computational complexity of image reconstruction.
  • Versatility: The method can be applied to various image reconstruction tasks, including computer graphics, animation, and image super-resolution.


Original Abstract Submitted

An image reconstruction method and apparatus are provided. An image reconstruction method includes determining an image warping result by warping a previous reconstruction result using change-data corresponding to a difference between rendered images, determining a previous filter kernel by executing a first neural network model with a previous rendered image and the image warping result, estimating a current filter kernel by warping the previous filter kernel using the change-data, and determining a current reconstruction result by executing a second neural network model with a current rendered image, the current filter kernel, and the image warping result.