17879307. PRODUCING A DEPTH MAP FROM A MONOCULAR TWO-DIMENSIONAL IMAGE simplified abstract (TOYOTA JIDOSHA KABUSHIKI KAISHA)

From WikiPatents
Jump to navigation Jump to search

PRODUCING A DEPTH MAP FROM A MONOCULAR TWO-DIMENSIONAL IMAGE

Organization Name

TOYOTA JIDOSHA KABUSHIKI KAISHA

Inventor(s)

Vitor Guizilini of Santa Clara CA (US)

Rares A. Ambrus of San Francisco CA (US)

Dian Chen of Mountain View CA (US)

Adrien David Gaidon of Mountain View CA (US)

Sergey Zakharov of San Francisco CA (US)

PRODUCING A DEPTH MAP FROM A MONOCULAR TWO-DIMENSIONAL IMAGE - A simplified explanation of the abstract

This abstract first appeared for US patent application 17879307 titled 'PRODUCING A DEPTH MAP FROM A MONOCULAR TWO-DIMENSIONAL IMAGE

Simplified Explanation

The patent application describes a system for generating a depth map using a neural network.

  • The system includes a processor and memory to store the neural network.
  • The neural network consists of three modules: encoding, multi-frame feature matching, and decoding.
  • The encoding module encodes an image to produce single-frame features.
  • The multi-frame feature matching module processes the single-frame features to generate information.
  • The decoding module decodes the information to produce the depth map.
  • The training datasets used for the multi-frame feature matching and encoding/decoding modules can be different.

Potential Applications:

  • Depth mapping in computer vision applications.
  • 3D reconstruction in virtual reality and augmented reality systems.
  • Object detection and tracking in autonomous vehicles.

Problems Solved:

  • Accurate depth mapping from 2D images.
  • Efficient processing of multi-frame features.
  • Training neural networks with different datasets for different modules.

Benefits:

  • Improved accuracy in depth mapping.
  • Enhanced performance in computer vision tasks.
  • Flexibility in training different modules with specific datasets.


Original Abstract Submitted

A system for producing a depth map can include a processor and a memory. The memory can store a neural network. The neural network can include an encoding portion module, a multi-frame feature matching portion module, and a decoding portion module. The encoding portion module can include instructions that, when executed by the processor, cause the processor to encode an image to produce single-frame features. The multi-frame feature matching portion module can include instructions that, when executed by the processor, cause the processor to process the single-frame features to produce information. The decoding portion module can include instructions that, when executed by the processor, cause the processor to decode the information to produce the depth map. A first training dataset, used to train the multi-frame feature matching portion module, can be different from a second training dataset used to train the encoding portion module and the decoding portion module.

Retrieved from "http://wikipatents.org/index.php?title=17879307._PRODUCING_A_DEPTH_MAP_FROM_A_MONOCULAR_TWO-DIMENSIONAL_IMAGE_simplified_abstract_(TOYOTA_JIDOSHA_KABUSHIKI_KAISHA)&oldid=24072"