Face Coordinate Alignment for Cross-Resolution Image Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Smart terminals with limited storage and computing capabilities face challenges in face recognition accuracy due to differences in face data obtained from images of varying resolutions, leading to reduced precision in recognition results.
Innovation Solution
An image processing method that acquires face coordinates from images of different resolutions, calculates a magnification ratio, and compares these coordinates to perform face clustering, thereby improving the accuracy of image processing by aligning images in the same coordinate system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If face recognition is performed on compressed images with lower resolution, then processing speed and storage efficiency are improved, but recognition precision deteriorates
Solution Approach 1:
The patent performs face recognition on the original high-resolution image first to obtain accurate face coordinates, then uses these coordinates as a reference for subsequent clustering operations. This preliminary recognition ensures high precision while the actual clustering is performed on downsampled images for efficiency.
Solution Approach 2:
The patent applies face recognition to only the relevant face regions identified in the original image, rather than processing the entire high-resolution image at full detail. This partial processing approach maintains recognition precision for face areas while reducing overall computational burden.
2Adaptability or versatility
If face recognition is performed on images with different resolutions, then adaptability to various devices is improved, but measurement precision of face coordinates deteriorates
Solution Approach 1:
The patent changes the resolution parameter of images adaptively based on device capabilities. Original high-resolution images are downsampled to lower resolutions for devices with limited processing power, while the system maintains a reference set of face coordinates from the original resolution to ensure precision across all devices.
Solution Approach 2:
The patent introduces face coordinates obtained from the original high-resolution image as an intermediary reference. These reference coordinates mediate between the high-resolution source and various downsampled versions, allowing accurate face identification across different resolutions by comparing relative position relationships rather than absolute coordinates.
3Productivity
If simplified algorithms are used for recognition classification, then processing efficiency is improved, but recognition precision deteriorates
Solution Approach 1:
The patent segments the recognition process into two stages: first, a simplified algorithm performs initial face detection and coordinate extraction from downsampled images for efficiency; second, a more sophisticated clustering algorithm refines the results by comparing face coordinates across multiple images, achieving high precision without requiring the simplified algorithm to handle the entire recognition task alone.
Data Source
AI summary
The present disclosure relates to an image processing method and apparatus, and an electronic device. The method includes: acquiring a first face coordinate of a first image; acquiring a second face coordinate of a second image, in which, the first image has a different resolution from the second image, and an image size of the first image is greater than an image size of the second image; calculating a magnification ratio according to the image size of the first image and the image size of the second image, and calculating a second target face coordinate according to the magnification ratio and the second face coordinate; comparing the first face coordinate with the second target face coordinate to obtain a comparing result; and performing face clustering according to the comparing result.