18358857. EFFICIENT COST VOLUME PROCESSING WITHIN ITERATIVE PROCESS simplified abstract (QUALCOMM Incorporated)
Contents
EFFICIENT COST VOLUME PROCESSING WITHIN ITERATIVE PROCESS
Organization Name
Inventor(s)
Risheek Garrepalli of San Diego CA (US)
Rajeswaran Chockalingapuramravindran of San Diego CA (US)
Jisoo Jeong of San Diego CA (US)
Fatih Murat Porikli of San Diego CA (US)
EFFICIENT COST VOLUME PROCESSING WITHIN ITERATIVE PROCESS - A simplified explanation of the abstract
This abstract first appeared for US patent application 18358857 titled 'EFFICIENT COST VOLUME PROCESSING WITHIN ITERATIVE PROCESS
Simplified Explanation
The abstract describes a method for processing cost volumes across multiple stages by varying receptive fields and performing learning-based correspondence estimation tasks.
- Processing a single level cost volume across multiple stages
- Varying receptive field across each processing stage
- Performing learning-based correspondence estimation task
- Processing different resolution of cost volume at each stage
- Maintaining same neighborhood sampling radius
- Increasing resolution from first stage to later stage
- Varying neighborhood sampling radius at each stage
- Maintaining same resolution
- Task may include optical flow estimation or stereo estimation
---
- Potential Applications
- Computer vision
- Image processing
- Robotics
- Autonomous vehicles
- Problems Solved
- Improving accuracy of correspondence estimation tasks
- Enhancing performance of optical flow and stereo estimation
- Optimizing processing of cost volumes across multiple stages
- Benefits
- Increased efficiency in processing cost volumes
- Improved accuracy in correspondence estimation
- Enhanced performance in optical flow and stereo estimation tasks
Original Abstract Submitted
A processor-implemented method comprises processing a single level cost volume across multiple processing stages by varying a receptive field across each of the processing stages. The method also includes performing a learning-based correspondence estimation task based on the processing. The varying may include processing a different resolution of the cost volume at each processing stage while maintaining a same neighborhood sampling radius. The resolution may increase from a first processing stage to a later processing stage. The varying may also include varying a neighborhood sampling radius at each of the processing stages while maintaining a same resolution. The task may be optical flow estimation or stereo estimation.