Perspective Image Generation Using Inverse Coordinate Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating perspective images in mobile terminals are CPU-intensive due to the need for floating-point and trigonometric operations, making it difficult to achieve smooth visual perception with limited processing time, especially in low- to mid-end cellular phone systems.
Innovation Solution
A method and device using inverse transformation to reduce the number of calculated pixel points by determining coordinates in the original image corresponding to those in the perspective image, with a preset function group that simplifies calculations and avoids floating-point operations, thereby improving generation speed and saving CPU resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If floating-point and trigonometric function operations are used for perspective transformation, then the accuracy of coordinate mapping is improved, but the CPU processing time and computational complexity increase significantly
Solution Approach 1:
The patent transforms the coordinate mapping function by changing parameters from floating-point trigonometric operations to integer-based polynomial operations. Specifically, it uses a polynomial approximation of the perspective transformation function that can be computed using only integer addition and multiplication, thereby maintaining sufficient accuracy while dramatically reducing computational complexity and processing time.
Solution Approach 2:
The patent replaces the computational mechanism of floating-point arithmetic and trigonometric function evaluation with a simpler integer-based polynomial computation system. This substitution eliminates the need for complex CPU floating-point units and trigonometric lookup tables, using instead simple integer operations that are native to the processor architecture.
2Speed
If perspective transformation is performed for every frame at 15+ frames per second, then smooth visual perception is achieved, but the computational load on the CPU becomes unsustainable in low- to mid-end systems
Solution Approach 1:
The patent changes the computational parameters from floating-point trigonometric calculations to integer polynomial calculations, reducing the operation count per frame by approximately 10-20 times. This enables the system to sustain 15+ frames per second processing rate without overwhelming the CPU resources of low- to mid-end mobile devices.
Solution Approach 2:
The patent pre-calculates and stores the polynomial coefficients for the perspective transformation function based on the vanishing point position. These pre-computed integer coefficients are then reused for each frame's coordinate mapping, eliminating the need to recompute complex trigonometric values for every pixel in every frame.
3Area of stationary object
If the entire original image is processed for perspective transformation, then complete coverage of the perspective image is achieved, but the calculation area and processing time increase
Solution Approach 1:
The patent segments the processing area by identifying and utilizing the vanishing point position to determine the actual rectangular region of the original image that maps to the perspective image. Instead of processing the entire original image, only the relevant segmented portion containing the mapped content is processed, significantly reducing the number of pixels requiring transformation.
Solution Approach 2:
The patent applies partial action by processing only the necessary portion of the original image that corresponds to the visible perspective view. By calculating the bounding box of the transformed region based on vanishing point geometry, it processes exactly the required area without excessive computation on irrelevant image regions.
Data Source
AI summary
A method and device for generating a perspective image are provided, which are related to multimedia technology. The method comprises the following steps: acquiring the length, height and vanishing point of a required perspective image and the length and height of an original image; for each pixel point (x′, y′) in the perspective image, determining the coordinate (x, y) of a pixel point in the original image corresponding to said pixel point; copying the pixel point at the coordinate (x, y) in the original image to the position corresponding to the coordinate (x′, y′) in said perspective image. The invention utilizes a principle that perspective image is smaller than original image, and employs a form of inverse transform, thereby the calculated pixel points are reduced, and compared with forward calculation, the calculation is greatly reduced, which improves the speed of perspective image generating and saves central processing unit (CPU) resources.