AM-FM Motion Estimation for Video Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion estimation methods in digital videos face limitations in accuracy and processing time, especially when dealing with noisy videos and complex textures, which hinders their effectiveness in applications like security and medical imaging.
Innovation Solution
The method employs 2D amplitude-modulation frequency-modulation (AM-FM) models to estimate motion by deriving equations from consecutive frames, utilizing instantaneous amplitude and frequency components, and combines these with existing methods like Horn-Schunck and Lucas-Kanade for improved accuracy and efficiency, particularly leveraging GPU implementations for rapid processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard optical flow methods based on intensity are used for motion estimation, then the processing time is reduced, but the accuracy of motion estimation deteriorates especially for complex textures and noisy videos
Solution Approach 1:
The patent segments the motion estimation problem into two distinct components: amplitude-modulated (AM) components representing translational motion and frequency-modulated (FM) components representing local deformations. This segmentation allows each component to be processed using optimized algorithms, improving overall accuracy without proportionally increasing processing time.
Solution Approach 2:
The patent transforms the motion estimation problem from intensity-based 2D spatial analysis to a combined amplitude-frequency domain analysis. By introducing the frequency dimension through FM component estimation, the method captures complex texture motions that are invisible to traditional intensity-based approaches, significantly improving measurement precision.
2Measurement precision
If advanced motion estimation models are used to improve accuracy for complex textures, then the measurement precision improves, but the device complexity and processing time increase
Solution Approach 1:
The patent replaces complex iterative optimization mechanisms with direct analytical solutions. By formulating motion estimation as an amplitude and frequency demodulation problem, the method eliminates the need for complex variational frameworks and iterative minimization, reducing computational complexity while maintaining high accuracy.
Solution Approach 2:
The patent changes the fundamental parameters used for motion estimation from intensity values to amplitude and frequency parameters. This parameter transformation simplifies the mathematical model, allowing direct computation of motion fields without complex iterative processes, thereby reducing device complexity.
3Reliability
If intensity-based optical flow methods are used, then the processing speed is maintained, but the ability to handle complex textures and noisy videos deteriorates
Solution Approach 1:
The patent performs preliminary decomposition of the video signal into amplitude-modulated and frequency-modulated components before motion estimation. This preliminary action separates the signal into meaningful components that are inherently more robust to noise and better suited for capturing complex texture motions, improving reliability without sacrificing processing speed.
Solution Approach 2:
The patent introduces amplitude and frequency components as intermediary representations between the raw video signal and the final motion estimation. These intermediaries filter out noise while preserving essential motion information, enhancing robustness to noise and complex textures while maintaining computational efficiency.
Data Source
AI summary
An AM-FM representation is used to derive AM and FM based equations that can be applied to two consecutive frames in parallel to derive motion estimates. The multidimensional AM-FM representations provide general representations of non-stationary content in digital images. The AM-FM estimate captures single images and features of a video that can lead to different applications in image and video analysis, for example, computer-aided diagnosis in medical applications or monitoring micro-movements of rocky material in the pit slopes.