Object Tracking Across Imaging Regions Without Occlusion Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle with accurately determining object identity when occlusion occurs, as images with occlusion can be included in the selected cluster, leading to decreased accuracy.
Innovation Solution
An image processing device that tracks objects in different image sequences, detects feature amount changes, and estimates imaging surfaces to determine object identity without using occluded data, by extracting feature amounts from images with no change and comparing them across different imaging regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a cluster having the largest number of images is selected for object identification, then the quantity of data available for analysis is increased, but images with occlusion may be included in the selected cluster leading to decreased determination accuracy
Solution Approach 1:
The patent changes the selection parameter from 'largest number of images' to 'images with no feature amount change'. This parameter transformation allows the system to maintain a sufficient quantity of images for analysis while excluding occluded images that would degrade identification accuracy. The feature amount change detection serves as a filtering mechanism that preserves data quantity while improving data quality.
Solution Approach 2:
The patent applies local quality control by evaluating each image's feature amount stability individually within the cluster. Instead of treating all images uniformly, the system identifies and selects only those images where the feature amount remains consistent, thereby ensuring high local quality (accuracy) of the selected subset while maintaining overall sufficient data quantity for reliable object identification.
2Quantity of substance
If feature amounts are extracted from all images in a cluster, then the quantity of data for comparison is increased, but occluded images may introduce errors in object identity determination
Solution Approach 1:
The patent transforms the data selection criterion from comprehensive inclusion to selective extraction based on feature amount stability. By monitoring whether feature amounts change between consecutive images, the system identifies reliable data points (no change) versus unreliable ones (change detected), thereby extracting feature amounts only from trustworthy images while maintaining sufficient data quantity for robust object identity determination.
3Measurement precision
If occluded images are excluded from analysis, then determination accuracy is improved, but the quantity of available data for comparison is reduced
Solution Approach 1:
The patent performs preliminary filtering of images based on feature amount change detection before proceeding to object identification. By pre-identifying and excluding occluded images through feature stability analysis, the system ensures that only high-quality images are used in the subsequent comparison process. This preliminary action maintains high determination accuracy while preserving sufficient data quantity by excluding only the necessary problematic images.
Data Source
AI summary
An image processing device includes a detection unit, a feature amount generating unit, a tracking unit, a feature amount change detecting unit, a tracking history extracting unit, and a determination unit, so that when occlusion in a selected cluster has occurred, the image processing device can determine whether or not objects are identical without using data of an image in which the occlusion has occurred. An image in which the occlusion has occurred can be excluded by using the data of the image determined to have no change in the feature amount, whereby a decrease in accuracy of determining whether or not the objects are identical can be suppressed.