17812340. PARTIAL SUPERVISION IN SELF-SUPERVISED MONOCULAR DEPTH ESTIMATION simplified abstract (QUALCOMM Incorporated)

From WikiPatents
Jump to navigation Jump to search

PARTIAL SUPERVISION IN SELF-SUPERVISED MONOCULAR DEPTH ESTIMATION

Organization Name

QUALCOMM Incorporated

Inventor(s)

Amin Ansari of Federal Way WA (US)

Avdhut Joshi of San Marcos CA (US)

Gautam Sachdeva of San Diego CA (US)

Ahmed Kamel Sadex of San Diego CA (US)

PARTIAL SUPERVISION IN SELF-SUPERVISED MONOCULAR DEPTH ESTIMATION - A simplified explanation of the abstract

This abstract first appeared for US patent application 17812340 titled 'PARTIAL SUPERVISION IN SELF-SUPERVISED MONOCULAR DEPTH ESTIMATION

Simplified Explanation

The abstract of this patent application describes techniques for machine learning using a depth model. Here is a simplified explanation of the abstract:

  • The patent application describes a method for generating a depth output from an input image frame using a depth model.
  • The depth model is updated based on a depth loss, which is determined by comparing the depth output with an estimated ground truth for the input image frame.
  • A total loss for the depth model is determined based on the depth loss, and the depth model is updated using this total loss.
  • The updated depth model is then used to generate a new depth output.

Potential applications of this technology:

  • This technology can be used in various computer vision applications, such as object recognition, scene understanding, and augmented reality.
  • It can also be applied in robotics for tasks like navigation and object manipulation, where accurate depth perception is crucial.

Problems solved by this technology:

  • The technique addresses the challenge of generating accurate depth information from 2D images, which is essential for many computer vision tasks.
  • It solves the problem of training a depth model by incorporating an estimated ground truth, allowing for continuous improvement and better depth predictions.

Benefits of this technology:

  • The method provides a more accurate and reliable way to estimate depth from images, improving the performance of computer vision systems.
  • By updating the depth model based on the total loss, the technique enables continuous learning and adaptation to different environments and scenarios.
  • The technology can enhance the capabilities of various applications, leading to improved object recognition, scene understanding, and robotic perception.


Original Abstract Submitted

Certain aspects of the present disclosure provide techniques for machine learning. A depth output from a depth model is generated based on an input image frame. A depth loss for the depth model is determined based on the depth output and an estimated ground truth for the input image frame, the estimated ground truth comprising estimated depths for a set of pixels of the input image frame. A total loss for the depth model is determined based at least in part on the depth loss. The depth model is updated based on the total loss, and a new depth output, generated using the updated depth model, is output.

Retrieved from "http://wikipatents.org/index.php?title=17812340._PARTIAL_SUPERVISION_IN_SELF-SUPERVISED_MONOCULAR_DEPTH_ESTIMATION_simplified_abstract_(QUALCOMM_Incorporated)&oldid=19417"