18481050. SCALING FOR DEPTH ESTIMATION simplified abstract (QUALCOMM Incorporated)
Contents
- 1 SCALING FOR DEPTH ESTIMATION
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 SCALING FOR DEPTH ESTIMATION - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Unanswered Questions
- 1.11 Original Abstract Submitted
SCALING FOR DEPTH ESTIMATION
Organization Name
Inventor(s)
Yinhao Zhu of La Jolla CA (US)
Jisoo Jeong of San Diego CA (US)
Yunxiao Shi of San Diego CA (US)
Fatih Murat Porikli of San Diego CA (US)
SCALING FOR DEPTH ESTIMATION - A simplified explanation of the abstract
This abstract first appeared for US patent application 18481050 titled 'SCALING FOR DEPTH ESTIMATION
Simplified Explanation
The patent application describes a system and method for processing sensor data, specifically for predicting depth maps for images using machine learning and depth values obtained from a tracker.
- The process involves determining a predicted depth map for an image using a trained machine learning system.
- The predicted depth map includes a predicted depth value for each pixel of the image.
- Depth values for the image are obtained from a tracker, which determines the depth values based on feature points between frames.
- The predicted depth map is scaled using the obtained depth values to provide scale-correct depth prediction values.
Potential Applications
This technology can be applied in various fields such as:
- Autonomous driving for accurate depth perception.
- Augmented reality for realistic depth rendering.
- Robotics for object detection and navigation.
Problems Solved
This technology addresses the following issues:
- Inaccurate depth estimation in images.
- Limited depth information for certain pixels.
- Lack of scale-correct depth prediction values.
Benefits
The benefits of this technology include:
- Improved accuracy in depth prediction.
- Enhanced visual perception in applications.
- Better object recognition and tracking capabilities.
Potential Commercial Applications
Potential commercial applications of this technology include:
- Development of advanced driver assistance systems.
- Integration into virtual reality and gaming platforms.
- Implementation in surveillance and security systems.
Possible Prior Art
One possible prior art for this technology could be the use of traditional computer vision techniques for depth estimation in images.
Unanswered Questions
How does the machine learning system handle variations in lighting conditions when predicting depth maps?
The patent application does not provide specific details on how the machine learning system accounts for changes in lighting conditions during depth map prediction.
What is the computational complexity of the process compared to existing depth estimation methods?
The patent application does not discuss the computational efficiency of the proposed system in relation to other depth estimation techniques.
Original Abstract Submitted
Systems and techniques are provided for processing sensor data. For example, a process can include determining, using a trained machine learning system, a predicted depth map for an image, the predicted depth map including a respective predicted depth value for each pixel of the image. The process can further include obtaining depth values for the image, the depth values including depth values for less than all pixels of the image from a tracker configured to determine the depth values based on one or more feature points between frames. The process can further include scaling the predicted depth map for the image using and the depth values. The output of the process can be scale-correct depth prediction values.
