Image Processing for Perspective Projection Distortion Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for non-perspective projection images, such as fisheye and equirectangular images, struggle to achieve uniform image processing across all fields of view in Head-Mounted Display (HMD) systems, leading to distorted wide-angle regions, especially when transforming images into perspective projections.
Innovation Solution
An image processing apparatus and method that transforms non-perspective projection images into perspective projection images within distinct regions, applies image processing, and then transforms the results back into non-perspective projection images, allowing for independent processing and reducing distortion by using a CPU and GPU to perform operations like bilateral filtering and alpha channel mapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If image processing is applied directly to non-perspective projection images (fisheye/equirectangular images), then processing can be performed on the original image format, but the wide-angle regions end up being recorded greatly distorted with respect to the central portion
Solution Approach 1:
The non-perspective projection image is divided into multiple image regions (first image region and second image region) with different fields of view. Image processing is applied separately to each region after transforming them into perspective projection images, allowing different processing parameters to be used for different regions to reduce overall distortion
Solution Approach 2:
A perspective projection image is introduced as an intermediate representation. The non-perspective projection image is transformed into a perspective projection image, image processing is applied to the perspective projection image, and then it is transformed back. This intermediary step enables proper handling of wide-angle regions while maintaining processing effectiveness
2Manufacturing precision
If image processing is changed in accordance with projection method and field of view to achieve uniform results, then uniform image processing can be achieved across all fields of view, but the device complexity increases
Solution Approach 1:
Different image processing parameters are applied to different image regions based on their specific characteristics. The first image region (first field of view) and second image region (second field of view) receive tailored processing parameters, allowing each region to be optimized independently while maintaining overall uniformity
Solution Approach 2:
Image processing parameters are adjusted according to the field of view and region characteristics. By changing parameters locally for different regions rather than using a single global set of parameters, uniform processing results are achieved across all fields of view without requiring overly complex system architecture
Data Source
AI summary
An image processing apparatus comprises one or more memories and one or more processors. The one or more processors and the one or more memories are configured to input a non-perspective projection image; transform a respective image in each image region of a plurality image regions in the non-perspective projection image into a respective perspective projection image; and, for each of the respective perspective projection images, after applying image processing to the respective perspective projection image, transform the result of the image processing into a respective non-perspective projection image.