X-ray Image Processing via Kernel-Based Feature Region Windowing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies for X-ray images are slow, making real-time analysis challenging.
Innovation Solution
An image processing method that involves obtaining feature map data, extracting feature regions based on the size of a convolution kernel, performing windowing processing, and generating a windowed feature map to enhance processing speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional image processing methods are used, then processing accuracy is maintained, but processing speed is slow and real-time analysis is difficult
Solution Approach 1:
The patent divides the feature map into multiple feature regions and processes them through parallel convolution operations. The feature map is segmented into distinct regions that can be simultaneously processed by multiple convolution kernels, thereby increasing processing speed while maintaining accuracy.
Solution Approach 2:
The patent introduces a new dimension by applying windowing operations across multiple feature maps in a sequence. This creates an additional processing dimension where windowed feature maps are generated and processed in parallel, significantly accelerating the overall processing speed and enabling real-time analysis.
2Productivity
If feature map processing is performed without windowing, then processing simplicity is maintained, but processing speed is insufficient for real-time applications
Solution Approach 1:
The patent performs windowing operations as a preliminary step before the final convolution processing. By pre-processing the feature maps with windowing operations and generating windowed feature maps in advance, the system prepares data structures that enable faster subsequent processing, thereby improving overall productivity without significantly increasing complexity.
Data Source
AI summary
The present disclosure provides an image processing method, an image processing device, an electronic apparatus and a readable storage medium. The image processing method includes: obtaining feature map data of an input image; extracting a feature region in the feature map data in accordance with a size of a convolution kernel; performing windowing processing on the feature region; and obtaining a windowed feature map of the input image in accordance with the feature region obtained after the windowing processing.