Image Processing Apparatus Motion Tracking via Asymmetric Area Orientation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stabilization methods face challenges in tracking feature points across multiple processing areas, leading to increased likelihood of feature point loss and decreased motion vector detection performance when processing areas are arranged in the same direction as the object's movement.
Innovation Solution
The solution involves determining the movement direction of an object and setting multiple processing areas in a direction different from the movement direction, with a selection unit that chooses tracking points within each area and a tracking unit that tracks these points across subsequent images, and in case of tracking failure, selecting a new point from adjacent or overlapping selection areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If processing areas are arranged in the same direction as object movement, then feature point tracking can be performed across areas, but the likelihood of feature points moving into another processing area increases causing tracking failure
Solution Approach 1:
The patent applies asymmetry by arranging processing areas perpendicular to the object movement direction rather than parallel to it. This asymmetric arrangement relative to the movement direction prevents feature points from crossing into adjacent processing areas during tracking, thereby maintaining tracking reliability without increasing system complexity.
2Measurement precision
If feature points are extracted uniformly across the entire image, then coverage is improved, but the distribution becomes non-uniform leading to concentrated feature points in certain areas
Solution Approach 1:
The patent applies segmentation by dividing the image into multiple processing areas and extracting feature points within each area independently. This segmentation approach ensures uniform distribution of feature points across the entire image while maintaining simple extraction operations within each segmented region, avoiding the concentration problem that occurs with whole-image extraction.
3Measurement precision
If template matching is performed on the entire reference image, then motion vector detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent applies segmentation by dividing the reference image into multiple processing areas and performing template matching independently in each area. This reduces the search space for each matching operation compared to searching the entire reference image, thereby improving processing speed while maintaining motion vector detection accuracy through sufficient sampling across all areas.
Solution Approach 2:
The patent applies local quality by performing template matching with focused attention on local regions (processing areas) rather than treating the entire image uniformly. This allows for accurate motion vector detection within each local area while reducing overall computational burden, achieving a balance between accuracy and processing speed.
Data Source
AI summary
There is provided an image processing apparatus. A determination unit determines a movement direction of an object. A setting unit sets, within a shooting range, a plurality of processing areas that are arranged in a different direction from the movement direction. A selection unit selects a tracking point in each processing area of a predetermined shot image. A tracking unit tracks, inside each processing area, the tracking point across one or more shot images that are shot after the predetermined shot image.