Keypoint Detection Circuit for Image Pyramid Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing pipelines consume significant CPU bandwidth and power when executing image processing algorithms, necessitating a hardware-based solution for efficient image processing, particularly for keypoint detection in image frames.
Innovation Solution
A keypoint detection circuit that generates an image pyramid by blurring and subsampling image data, using multiple branches to determine multiple sets of keypoints at different blur levels, allowing for a larger and more varied set of keypoints for object detection and matching across images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If image processing algorithms are executed on CPU, then image processing can be performed, but CPU bandwidth consumption increases and power consumption increases
Solution Approach 1:
The patent replaces CPU-based software execution with a dedicated hardware image processing pipeline that includes specialized circuits for keypoint detection, filtering, and matching operations. This hardware implementation performs image processing functions independently from the CPU, eliminating the need for CPU bandwidth while maintaining full image processing capability.
2Ease of operation
If image processing algorithms are executed on CPU, then image processing can be performed, but power consumption increases
Solution Approach 1:
The patent implements a dedicated hardware image processing pipeline with specialized circuits for keypoint detection, filtering, and matching operations. This hardware implementation performs image processing functions independently from the CPU, eliminating the need for CPU bandwidth while maintaining full image processing capability.
3Measurement precision
If multiple sets of keypoints are determined for each octave of the pyramid using different levels of blur, then a larger and more varied set of keypoints is obtained, but device complexity increases
Solution Approach 1:
The patent divides the image processing task into multiple parallel branches, where each branch is configured to determine keypoints at a specific blur level. This segmentation allows the system to process different blur levels simultaneously through dedicated circuits in each branch, achieving high keypoint detection accuracy while managing complexity through modular parallel architecture.
Solution Approach 2:
The patent extends keypoint detection into the blur level dimension by creating multiple branches that operate at different blur levels (e.g., unblurred, first blur level, second blur level). This dimensional expansion allows the system to detect keypoints across multiple scales and blur conditions, significantly increasing the variety and accuracy of detected keypoints without requiring sequential processing.
Data Source
AI summary
Embodiments relate a keypoint detection circuit for identifying keypoints in captured image frames. The keypoint detection circuit generates an image pyramid based upon a received image frame, and determine multiple sets of keypoints for each octave of the pyramid using different levels of blur. In some embodiments, the keypoint detection circuit includes multiple branches, each branch made up of one or more circuits for determining a different set of keypoints from the image, or for determining a subsampled image for a subsequent octave of the pyramid. By determining multiple sets of keypoints for each of a plurality of pyramid octaves, a larger, more varied set of keypoints can be obtained and used for object detection and matching between images.